JP2018526362A - Biomarkers for the treatment of alopecia areata - Google Patents

Abstract

  The subject matter of this disclosure relates to improved diagnosis and prognosis of alopecia areata and the ability to identify patient subpopulations that respond to such treatment with respect to biomarkers that enable effective treatment for this disease. Methods of incorporating biomarkers with and methods of incorporating biomarkers with the ability to track the progress of such treatment.

Description

This application claims priority based on US Provisional Patent Application No. 62 / 205,476, filed Aug. 14, 2015, the entire contents of which are hereby incorporated by reference. .

Grant Information This invention was made with government support under NIH Grant Numbers R01AR056016, R21AR061881 and 5U01AR067173 awarded by the National Institutes of Health. The government has certain rights in this invention.

INTRODUCTION The subject matter disclosed herein relates to improved diagnosis and prognosis of alopecia areata and biomarkers that enable effective treatment for this disease, and patient subpopulations that respond to such treatment Methods of incorporating biomarkers with the ability to identify and methods of incorporating biomarkers with the ability to track the progress of such treatment.

  Alopecia areata (AA) is an autoimmune skin disease in which the hair follicle is the target of an immune attack. Patients typically exhibit a round or oval hair loss portion on the scalp that can naturally dissipate, persist, or progress to involve the scalp or the entire body. The three major phenotypic variants of the disease are patchy AA (AAP), often located in small oval regions of the scalp or within the region of the beard, total alopecia (AT) involving the entire scalp ) And systemic alopecia (AU) involving the entire body surface.

  Currently, there are no FDA-approved drugs for AA and treatment is often empirical, but typically, observation, intralesional steroids, topical immunotherapy or unproven efficacious broad immunity With suppression treatment. AU and AT, which are more severe forms of the disease, are often refractory to treatment. Furthermore, a common assumption among dermatologists and therapists is that long-term AU and AT become unrecoverable or change the scalp to a “burned out” condition, which depends on the duration and treatment of the disease. This is supported by the inverse correlation of responsiveness to. Despite its high morbidity, there remains a need for biomarkers that identify the severity of the disease, as well as effective treatment for the disease, and the ability to identify patient subpopulations that respond to such treatment. There remains a need for methods of incorporating biomarkers with and for incorporating biomarkers with the ability to track the progress of such treatments.

  The present disclosure relates to biomarkers that allow improved diagnosis and prognosis of alopecia areata and effective treatment for disease. In certain embodiments, a method for treating alopecia areata (AA) in a subject identifies the subject's AA disease severity by detecting a biomarker indicative of the disease severity, and Providing the subject with a therapeutic intervention appropriate to the identified disease severity. The subject matter of the present disclosure further provides a method for treating AA in a subject, the method comprising detecting a biomarker indicative of the tendency to respond to JAK inhibitor treatment in a subject having AA. Identifying and administering a JAK inhibitor to the subject when the identified biomarker indicates a tendency for the subject to respond to the inhibitor. The subject matter of the present disclosure further provides a method of treating alopecia areata (AA) in a subject, the method comprising administering a JAK inhibitor to a subject having AA; and responsiveness to the JAK inhibitor treatment. Detecting a biomarker indicative of

  In certain embodiments, the detection of the biomarkers of the present disclosure described above is performed on a sample obtained from a subject, the sample being skin, blood, serum, plasma, urine, saliva, sputum, mucus, semen , Selected from the group consisting of amniotic fluid, mouth washes and bronchial washes. In certain embodiments, the subject is a human. In certain embodiments, the sample is a skin sample. In certain embodiments, the sample is a serum sample. In certain embodiments, the biomarker is a gene expression signature. In certain embodiments, the gene expression signature includes one or more gene expression information from the following group of genes: KRT-related genes; CTL-related genes; and IFN-related genes. In certain embodiments, the KRT-related genes include DSG4, HOXC31, KRT31, KRT32, KRT33B, KRT82, PKP1 and PKP2. In certain embodiments, CTL related genes include CD8A, GZMB, ICOS and PRF1. In certain embodiments, the IFN-related genes include CXCL9, CXCL10, CXCL11, STAT1 and MX1. In certain embodiments, the gene expression signature is alopecia areata disease activity index (ALADIN). In certain embodiments, the gene expression signature is an alopecia areata gene signature (AAGS) comprising one or more genes as defined in Table A. In certain embodiments, the gene expression signature is IKZF1, DLX4, or a combination thereof.

  In certain embodiments, detection of a biomarker of the present disclosure is performed by a nucleic acid hybridization assay. In certain embodiments, detection is performed by microarray analysis. In certain embodiments, detection is performed by polymerase chain reaction (PCR) or nucleic acid sequencing. In certain embodiments, the biomarker is a protein. In certain embodiments, the presence of the protein is detected using a reagent that specifically binds to the protein. In certain embodiments, the reagent is a monoclonal antibody or antigen-binding fragment thereof, or a polyclonal antibody or antigen-binding fragment thereof. In certain embodiments, detection is performed by enzyme linked immunosorbent assay (ELISA), immunofluorescence measurement or Western blot assay.

  In another aspect, the presently disclosed subject matter provides a kit for treating alopecia areata (AA) in a subject, which kit is useful for detecting a biomarker indicative of a subject's disease severity. One or more detection reagents and one or more treatment reagents useful for treating AA. The subject matter of the present disclosure may further provide a kit for treating alopecia areata (AA) in a subject, the kit being one or more useful for treating AA in a subject. It includes one or more detection reagents useful for detecting biomarkers that exhibit a tendency to respond to the above treatment reagents and one or more treatment reagents useful for treating AA. In certain embodiments, the kit further comprises one or more probe sets, arrays / microarrays, biomarker specific antibodies and / or beads. In certain embodiments, the kit further comprises instructions. In certain embodiments, the treatment reagent may be selected from JAK inhibitors.

  In certain embodiments, the JAK inhibitor is a Jak1 / Jak2 / Jak3 / Tyk2 / STAT1 / STAT2 / STAT3 / STAT4 / STAT5a / STAT5b / STAT6 / OSM / gp130 / LIFR / OSM-Rβ gene, or a Jak1 / Jak2 / It is a compound that interacts with Jak3 / Tyk2 / STAT1 / STAT2 / STAT3 / STAT4 / STAT5a / STAT5b / STAT6 / OSM / gp130 / LIFR / OSM-Rβ protein. In certain embodiments, the JAK inhibitor is ruxolitinib (INCB 018424), tofacitinib (CP690550), tyrophostin AG490 (CAS Number: 133550-30-8), thimerotinib (CYT387), paclitinib (SB1518), valicinib (SB9518), (TG101348), BMS-911543 (CAS Number: 1271022-90-2), Restaurtinib (CEP-701), Fludarabine, Epigallocatechin-3-gallate (EGCG), Peficitinib, ABT494 (CAS Number: 131266-2-60-3) ), AT9283 (CAS Number: 896466-04-9), Desernotinib, Filgotinib, Gand Tinib, INCB39110 (CAS Number: 1334298-90-6), PF 04965842 (CAS Number: 1629022-68-4), R348 (R-932348, CAS Number: 916742-11-5; 1620142-65-5), AZD1480 (CAS Number: 935666-88-9), Serdulatinib, INCB052793, NS018 (CAS Number: 1239358-86-1 (free base); 1239358-85-0 (HCl)), AC 410 (CAS Number: 1361415- 84-0 (free base), 136164-15-86-2 (HCl)), CT1578 (SB1578, CAS Number: 937273-04-6), JTE052, P 6263276, R548, TG02 (SB1317, CAS Number: 937270-47-8), Lambricus revelus extract, ARN4079, AR13154, UR67767, CS510, VR588, DNX04042, hyperforin, its derivatives, its deuterated variants, It may be selected from its salts or combinations thereof. In certain embodiments, the detection reagent may be selected from fluorescent reagents, luminescent reagents, dyes, radioisotopes, derivatives thereof or combinations thereof.

  The detailed description, given by way of example and not intended to limit the invention to the specific embodiments described, can be understood in conjunction with the appended drawings.

50 shows a specific signature for alopecia areata disease, the 50 most differentially expressed genes with increased expression in AA specific disease signatures between AT / AU, AAP and NC samples in the training set, and , A heat map of the 50 most differentially expressed genes with reduced expression. FIG. 3 is a diagram showing signatures specific to alopecia areata disease, and is an expression topographic map of a sample aligned along the main components of differential gene expression. The dots represent the position of each sample in the expression space (white = NC, blue = AAP, red = AT / AU), and the peak size is generated based on the number of samples in the region (more The peaks produced by the juxtaposed sample are higher and wider). The principal component space can be condensed into a single numerical score that reflects the risk that the sample is control, AAP or AT / AU based on its position in the terrain space. This consensus score provides a sample cohort of statistically significant fractional controls, AAP and AT / AU (box plot). Boxes indicate interquartile range and median, whiskers indicate percentiles of 5 and 95, * indicates statistical significance for NC, and † indicates statistical significance for AAP. FIG. 4 is a diagram of increased gene expression complexity and persistent inflammation in total head alopecia and systemic alopecia, AT / AU compared to normal (“AT / AU”), and AAP compared to normal (“AAP” ") Is a Venn diagram of genes having differential expression. The number of genes with differential expression within each section of the Venn diagram is shown. FIG. 4 is a diagram of increased gene expression complexity and persistent inflammation in total head alopecia and systemic alopecia, and peri-follicular / sphere of CD3 infiltrates between skin sections from patients with AT / AU, AAP or NC Histopathological score of the range. * P <0.01; ** p <0.0005 Diagram of increased gene expression complexity and persistent inflammation in total head alopecia and systemic alopecia, with representative histological images reflecting HPS score 0 (no infiltration) to 3 (severe infiltration) is there. FIG. 4 is a diagram of increased gene expression complexity and persistent inflammation in total head alopecia and systemic alopecia, a list of KEGG pathways shared between AT / AU versus normal controls and AAP versus normal controls. Schematic illustration of increased gene expression complexity and persistent inflammation in global alopecia and systemic alopecia, with KEGG pathway up-regulated in AT / AU vs. normal control (red), AAP vs. normal control (blue) Or a network map of shared paths in both AT / AU vs. normal control and AAP vs. normal control. Gene expression analysis that occurs between individuals in AA, comparing patient-matched lesion and non-lesion samples, identifying them to each other and to genes that represent healthy controls. Gene expression analysis that occurs between individuals in AA, comparing patient-matched lesion and non-lesion samples, identifying them to each other and to genes that represent healthy controls. Gene expression analysis that occurs between individuals in AA, comparing patient-matched lesion and non-lesion samples, identifying them to each other and to genes that represent healthy controls. Gene expression analysis occurring between individuals in AA, with lesion (red) and non-lesion (blue) samples matched patients aligned by normalized deviation from the first principal component of expression difference is there. The length of the line between the paired samples is based on the consensus of all signature genes, indicating overall similarity (shorter lines) or dissimilarity (longer lines). Gene expression analysis that occurs between individuals in AA, using the first two principal component analysis of normal control sample, lesion AAP and non-lesion AAP sample, the display of the lesion sample is not any cohort It has been shown that clusters form between lesion samples and controls. Gene expression analysis occurring between individuals in AA, analysis of overexpressed and underexpressed gene pathways and functional annotation between samples of lesioned AAP and non-lesioned AAP up in non-lesion and in lesion It reveals a separate set of genes that goes down (red nodes) and goes down in non-lesions and up in lesions (blue nodes). These genes are associated with the indicated enrichment annotation (yellow nodes) and represent functional molecular differences between the lesioned and non-lesioned samples reflected in their expression profiles. FIG. 5 shows immune cell invasion gene expression signatures that correlate with AA phenotype and is a relative estimate of infiltrated immune cells shown per patient based on consensus expression of corresponding immune markers (heat map, right ). Increasing red color indicates increasing amount of infiltrate. Patients are ranked by CD8 infiltration. Ranking patients with the highest to lowest CD8 infiltration reveals population bias. The boxplot reflects the distribution of the presented clinical presentation according to the CD8 infiltration rank (the lower the rank, the higher the level of infiltration shown). Box indicates quartile range and median, whiskers indicate percentiles of 5 and 95, * indicates statistical significance p = 0.005, and ** indicates p <1 × 10−5. FIG. 5 shows immune cell invasion gene expression signatures that correlate with AA phenotype, using consensus expression, the invasion contamination of biopsy samples is shown in each presentation cohort (AAP = mottled, (left pie), AT / AU = Total head alopecia / systemic alopecia (right pie)) and estimated relative share of each immune tissue type (line graph) in total infiltrating contaminants compared to unaffected controls. FIG. 5 shows immune cell invasion gene expression signatures that correlate with AA phenotype, in estimated invasion of each indicated immune type, expressed as a percentage of total sample signal across NC, AAP and AT / AU. It is a change. FIG. 6 shows an ALADIN score corresponding to a disease phenotype, and simultaneous expression analysis of genes with expression differences between AA and healthy controls reveals 20 module genes. FIG. 5 shows the ALADIN score corresponding to the disease phenotype, with GSEA of all 20 gene modules for enrichment in significant expression differences between AA and controls, with the green and brown modules being the most highly enriched in comparison It has been made clear. Figure 2 shows the ALADIN score corresponding to the disease phenotype, and pathway analysis of these two modules (circles) revealed significant enrichment of several immune and immune response pathways (orange diamonds) Includes previously associated genes in GWAS (yellow), ALADIN CTL gene (magenta), CTL gene that is also a GWAS hit (pink), and ALADIN IFN gene (turquoise). FIG. 5 shows an ALADIN score corresponding to a disease phenotype, which is a three-dimensional integrated integrated structural change reflected by immune invasion and gene expression to identify the relative risk of AA severity in a patient Classify patient samples (black: NC, green: AAP, red: AT / AU). FIG. 5 shows the ALADIN score corresponding to the disease phenotype, with CTL (top panel), IFN (middle panel) and KRT (bottom panel) signature scores from patients with AU / AT for disease duration. Show. It is the figure which showed the AA verification set, and has shown the dendrogram and heat map of 33 samples in a verification data set. Hierarchical clustering using Euclidean distances and mean linkages was used to identify 2002 Affymetrix PSIDs identified as differentially expressed between AA patients and normal controls in the Discovery data set used to cluster the samples. Done using. FIG. 7 shows T cell immunity gene signatures between AA samples, where unsupervised consensus clustering of AA patients and unaffected controls using signature genes unique to each invasive immune tissue shows relative infiltrates in each sample. Quantification is possible. In this heat map, red indicates higher expression and white indicates lower expression. The boundaries of the three major superclusters are defined as relative levels of low, medium and high invasion based on marker expression. FIG. 3 shows T cell immunity gene signatures between AA samples, with three invasion superclusters statistically significantly correlated with prognosis. A 3x3 chi-square test reveals that the severity of invasion predicts the severity of the AA phenotype across these patients. The number displayed in each cell represents the percentage of each clinical presentation found in the accompanying supercluster, for example, 72% of NC samples were found in low clusters. Chi-square statistics and associated p-values are provided. FIG. 4 shows that the module in the AA disease-specific signature defines the ALADIN component, and the dendrogram reflects the gene co-expression clustering results. Along the bottom, a colored barcode indicates the compartment that identified the 20 co-expressed modules used in this study. FIG. 5 shows that modules in AA disease-specific signatures define ALADIN components, showing a table of results when testing several clinical traits for association with 20 modules. In each cell, a p-value for the association between the corresponding module and the character is displayed. The cells are colored red that increases to correspond to the significance of the association. FIG. 5 shows that modules in AA disease-specific signatures define ALADIN components, showing Gene Set Enrichment Analyzes that test the statistical enrichment of each of the original ALADIN pathways in the AA cohort. In all comparisons to unaffected controls, there was a statistical enrichment of genes in the IFN, CTL and KRT pathways in the expected direction (IFN and CTL were positively enriched and KRT was negatively enriched). . FIG. 5 shows that modules in AA disease-specific signatures define ALADIN components, showing Gene Set Enrichment Analyzes that test the statistical enrichment of each of the original ALADIN pathways in the AA cohort. In all comparisons to unaffected controls, there was a statistical enrichment of genes in the IFN, CTL and KRT pathways in the expected direction (IFN and CTL were positively enriched and KRT was negatively enriched). . FIG. 5 shows that modules in AA disease-specific signatures define ALADIN components, showing Gene Set Enrichment Analyzes that test the statistical enrichment of each of the original ALADIN pathways in the AA cohort. In all comparisons to unaffected controls, there was a statistical enrichment of genes in the IFN, CTL and KRT pathways in the expected direction (IFN and CTL were positively enriched and KRT was negatively enriched). . FIG. 6 shows that the ALADIN component distinguishes between AA phenotype and normal control, wherein the components of ALADIN CTL (upper panel), IFN (middle panel) and KRT (lower panel) are normal control, A comparison was made between AAP and AU / AT samples. * P <0.05; ** p <0.0001. The time period does not significantly affect the ALADIN component score among AAP patients. The ALADIN CTL (upper panel), IFN (middle panel) and KRT (lower panel) components Comparisons were made between AAP patients with a duration of less than a year or at least 5 years. No significant difference was found. FIG. 6 shows that CD8 + NKG2D + cytotoxic T lymphocytes accumulate in the skin and are necessary and sufficient to induce disease in AA mice, and are the inner root of the hair follicle (marked by K71) Figure 8 shows immunofluorescent staining of NKG2D ligand (H60) in the sheath. Scale bar, 100 μm. FIG. 6 shows that CD8 + NKG2D + cytotoxic T lymphocytes accumulate in the skin and are necessary and sufficient to induce disease in AA mice, C57BL / 6, healthy C3H / HeJ and C3H / HeJAA mice CD8 + NKG2D + cells in the hair follicle of FIG. Upper scale bar, 100 μm; lower scale bar, 50 μm. FIG. 7 shows that CD8 + NKG2D + cytotoxic T lymphocytes accumulate in the skin and are necessary and sufficient to induce disease in AA mice, cutaneous lymphadenopathy and cell hyperplasia in C3H / HeJAA mice Is shown. FIG. 4 shows that CD8 + NKG2D + cytotoxic T lymphocytes accumulate in the skin and are necessary and sufficient to induce disease in AA mice, and the skin and skin draining lymphs of alopecia versus unimplanted mice The frequency of CD8 + NKG2D + T cells in the node (number shown above box area) is shown. FIG. 6 shows that CD8 + NKG2D + cytotoxic T lymphocytes accumulate in the skin and are necessary and sufficient to induce disease in AA mice, and CD8 + NKG2D + T cells in the skin lymph nodes of C3H / HeJ alopecia mice Shows the immune phenotype. FIG. 2 shows that CD8 + NKG2D + cytotoxic T lymphocytes accumulate in the skin and are necessary and sufficient to induce disease in AA mice, with Rae-lt expression grown from C3H / HeJ hair follicles Shows dose-dependent specific cell lysis (right) induced by CD8 + NKG2D + T cells isolated from AA mouse skin lymph nodes in the presence of dermal sheath cells (left), blocking anti-NKG2D antibody or isotype control ing. Effector target ratios are given as shown. Data are mean ± s. d. It is expressed as FIG. 4 shows that CD8 + NKG2D + cytotoxic T lymphocytes accumulate in the skin and are necessary and sufficient to induce disease in AA mice, all lymph node (LN) cells, CD8 + NKG2D + T cells only, CD + NKG2D -Shows hair loss in C3H / HeJ mice (5 mice per group) injected subcutaneously with lymph node cells depleted of T cells or NKG2D +. Mice are representative of two experiments. *** P <0.001 (Fisher exact test). FIG. 6 shows prevention of AA by a blocking antibody against IFN-γ, IL-2 or IL-15Rβ, and shows that C3H / HeJ transplanted mice were treated systemically from the time of transplantation, and antibodies against IFN-γ 2 shows the development of AA in C3H / HeJ transplanted mice treated systemically from the time of transplantation using. FIG. 6 shows prevention of AA by a blocking antibody against IFN-γ, IL-2 or IL-15Rβ, and shows that C3H / HeJ transplanted mice were treated systemically from the time of transplantation, and antibodies against IFN-γ 2 shows the development of AA in C3H / HeJ transplanted mice treated systemically from the time of transplantation using. FIG. 6 shows the frequency of CD8 + NKG2D + T cells in the skin of mice treated with an antibody to IFN-γ compared to PBS-treated mice (numbers shown above the box area). (** P <0.05, ** P <0.01, *** p <0.001) It is the figure which showed the prevention of AA by the blocking antibody with respect to IFN-gamma, IL-2, or IL-15R (beta), and showed that the C3H / HeJ transplant mouse | mouth was treated systemically from the time of a transplant, using an isotype control antibody. Or immunohistochemical staining of skin biopsies showing CD8 and MHC class I and II expression in the skin of mice treated with antibodies to IFN-γ. Scale bar, 100 μm. FIG. 6 shows prevention of AA by blocking antibodies against IFN-γ, IL-2 or IL-15Rβ, and shows that C3H / HeJ transplanted mice were treated systemically from the time of transplantation, and antibodies against IL-2 2 shows the development of AA in C3H / HeJ transplanted mice treated systemically from the time of transplantation using. FIG. 6 shows prevention of AA by blocking antibodies against IFN-γ, IL-2 or IL-15Rβ, and shows that C3H / HeJ transplanted mice were treated systemically from the time of transplantation, and antibodies against IL-2 2 shows the development of AA in C3H / HeJ transplanted mice treated systemically from the time of transplantation using. FIG. 6 shows the frequency of CD8 + NKG2D + T cells in the skin of mice treated with an antibody to IL-2 compared to PBS-treated mice (number shown above box area). (** P <0.05, ** P <0.01, *** p <0.001) It is the figure which showed the prevention of AA by the blocking antibody with respect to IFN-gamma, IL-2, or IL-15R (beta), and showed that the C3H / HeJ transplant mouse | mouth was treated systemically from the time of a transplant, using an isotype control antibody. Or immunohistochemical staining of skin biopsies showing CD8 and MHC class I and II expression in the skin of mice treated with antibodies to IL-2. Scale bar, 100 μm. FIG. 2 shows prevention of AA by blocking antibodies against IFN-γ, IL-2 or IL-15Rβ, and shows that C3H / HeJ transplanted mice were treated systemically from the time of transplantation, and antibodies against IL-15Rβ 2 shows the development of AA in C3H / HeJ transplanted mice treated systemically from the time of transplantation using. FIG. 2 shows prevention of AA by blocking antibodies against IFN-γ, IL-2 or IL-15Rβ, and shows that C3H / HeJ transplanted mice were treated systemically from the time of transplantation, and antibodies against IL-15Rβ 2 shows the development of AA in C3H / HeJ transplanted mice treated systemically from the time of transplantation using. FIG. 6 shows the frequency of CD8 + NKG2D + T cells in the skin of mice treated with antibodies to IL-15Rβ (numbers shown above the box area) compared to PBS-treated mice. (** P <0.05, ** P <0.01, *** p <0.001) It is the figure which showed the prevention of AA by the blocking antibody with respect to IFN-gamma, IL-2, or IL-15R (beta), and showed that the C3H / HeJ transplant mouse | mouth was treated systemically from the time of a transplant, using an isotype control antibody. Or immunohistochemical staining of skin biopsies showing CD8 and MHC class I and II expression in the skin of mice treated with antibodies to IL-15Rβ. Scale bar, 100 μm. Suppression of systemic JAK1 / 2 or JAK3 is shown to prevent the development of AA in transplanted C3H / HeJ mice, and C3H / C treated systemically from the time of transplantation with luxolitinib (JAK1 / 2i) Fig. 4 shows the occurrence of AA in HeJ transplanted mice. (** P <0.01). Hair regeneration is also shown 12 weeks after discontinuation of treatment. Scale bar, 100 μm. Suppression of systemic JAK1 / 2 or JAK3 is shown to prevent the development of AA in transplanted C3H / HeJ mice, and C3H / C treated systemically from the time of transplantation with luxolitinib (JAK1 / 2i) Fig. 4 shows the occurrence of AA in HeJ transplanted mice. (** P <0.01) FIG. 4 shows that suppression of systemic JAK1 / 2 or JAK3 prevents the development of AA in transplanted C3H / HeJ mice, and the skin and skin of mice treated with PBS or with JAK1 / 2i The frequency of CD8 + NKG2D + T cells in the lymph nodes (number shown above box area) is shown. (*** p <0.001) Suppression of systemic JAK1 / 2 or JAK3 prevents the development of AA in transplanted C3H / HeJ mice, and CD8 in the skin of mice treated with PBS or with JAK1 / 2i FIG. 6 shows immunohistochemical staining of skin biopsies showing MHC class I and II expression. Inhibition of systemic JAK1 / 2 or JAK3 is shown to prevent the development of AA in transplanted C3H / HeJ mice, given as log2 mean expression Z-score, using PBS, or JAK1 / 2 shows the ALADIN score of transcriptional analysis from mice treated with 2i. Suppression of systemic JAK1 / 2 or JAK3 shows that AA development is prevented in transplanted C3H / HeJ mice, and C3H / HeJ transplant treated systemically from the time of transplantation with tofacitinib (JAK3i) 2 shows the occurrence of AA in mice. (** P <0.01). Hair regeneration is also shown 12 weeks after discontinuation of treatment. Scale bar, 100 μm. Suppression of systemic JAK1 / 2 or JAK3 shows that AA development is prevented in transplanted C3H / HeJ mice, and C3H / HeJ transplant treated systemically from the time of transplantation with tofacitinib (JAK3i) 2 shows the occurrence of AA in mice. (** P <0.01) FIG. 3 shows that suppression of systemic JAK1 / 2 or JAK3 prevents the development of AA in transplanted C3H / HeJ mice, and the skin and lymph nodes of mice treated with PBS or with JAK3i Shows the frequency of CD8 + NKG2D + T cells (numbers shown above the box area). (*** p <0.001) Suppression of systemic JAK1 / 2 or JAK3 is shown to prevent the development of AA in transplanted C3H / HeJ mice, CD8 and MHC in the skin of mice treated with PBS or with JAK3i FIG. 2 shows immunohistochemical staining of skin biopsies showing class I and II expression. Suppression of systemic JAK1 / 2 or JAK3 is shown to prevent the development of AA in transplanted C3H / HeJ mice, given as log2 mean expression Z-score, using PBS, or JAK3i 2 shows the ALADIN score of transcription analysis from mice treated with FIG. 7 shows the recovery of established AA using a local small molecule inhibitor of downstream effector kinase JAK1 / 2 or JAK3 and the clinical outcome of patients with AA, with 0 daily administration for 12 weeks One group with AA over a long period (at least 12 weeks post-transplantation) with the back treated locally with 5% JAK1 / 2i (middle), 0.5% JAK3i (bottom) or vehicle alone (Aquaphor, top) Three mice are shown. This experiment was repeated three times. An additional 8 weeks of hair regeneration after discontinuation of treatment is also shown. FIG. 4 shows the recovery of established AA using local small molecule inhibitors of downstream effector kinase JAK1 / 2 or JAK3 and the clinical outcome of patients with AA, shown as the week after treatment It shows the time course of the hair growth index. FIG. 4 shows the recovery of established AA using local small molecule inhibitors of downstream effector kinase JAK1 / 2 or JAK3 and the clinical outcome of patients with AA, compared to solvent control mice, JAK1 Shown is the frequency of CD8 + NKG2D + T cells in the skin of mice treated with / 2i or JAK3i (number shown above the box area) (mean ± sem, n = per group) 3, * P <0.05, ** P <0.01). NS, not significant. FIG. 2 shows the recovery of established AA using local small molecule inhibitors of downstream effector kinase JAK1 / 2 or JAK3 and the clinical outcome of patients with AA, where the ALADIN score is log2 mean expression Z -Shows treatment-related reductions in CTL and IFN signatures given as scores. FIG. 2 shows recovery of established AA using a local small molecule inhibitor of downstream effector kinase JAK1 / 2 or JAK3, and clinical results of a patient with AA, immune tissue of a mouse skin biopsy Chemical staining shows a treatment-related decrease in the expression of CD8 and MHC class I and II markers. Scale bar, 100 μm. FIG. 4 shows recovery of established AA using local small molecule inhibitors of downstream effector kinase JAK1 / 2 or JAK3 and clinical results of patients with AA, using baseline with ruxolitinib Shown is treatment of patient 3 with AA having> 80% hair loss of the scalp and hair regeneration 12 weeks after oral treatment. FIG. 4 shows the recovery of established AA using a local small molecule inhibitor of downstream effector kinase JAK1 / 2 or JAK3 and the clinical outcome of a patient with AA, before treatment of patient 2 (baseline ) And a clinical correlation study of biopsies obtained 12 weeks after treatment, class I (A, B, C) and class II (DP) of CD4, CD8 and human leukocyte antigen (HLA) , DQ, DR). Scale bar, 200 μm. FIG. 6 shows recovery of established AA using local small molecule inhibitors of downstream effector kinases JAK1 / 2 or JAK3 and clinical results of patients with AA, showing AA presented as a heat map Figure 2 shows RNA microarray analysis from treated patients 1 and 2 (pre-treatment versus post-treatment versus 3 normal subjects). KRT, hair follicle keratin. FIG. 6 shows the recovery of established AA using local small molecule inhibitors of downstream effector kinase JAK1 / 2 or JAK3 and the clinical outcome of patients with AA, presented as a cumulative ALADIN index Figure 2 shows RNA microarray analysis from treated patients 1 and 2 with (pre-treatment versus post-treatment versus 3 normal subjects). KRT, hair follicle keratin. FIG. 2 shows a clinical photograph and ALADIN profile of serum CXCL10 of a scalp skin biopsy sample from an AA patient treated with tofacitinib. In the upper panel, photographs of the posterior scalp were obtained over the course of treatment with 5 mg of tofacitinib twice daily for 16 weeks. In the lower left panel, blood and scalp skin samples were taken 4 weeks after treatment with baseline and tofacitinib. A CXCL10 ELISA was performed. The lower center panel shows a heat map of ALADIN gene from healthy control patients (normal) and from scalp skin samples taken from AA patients at baseline (T0) and 4 weeks after treatment (T4). Has been. In the lower right panel, ALADIN plots of scalp skin samples taken from healthy control patients (black) and from AA patients at baseline (red) and 4 weeks after treatment (yellow) are shown. FIG. 5 shows recurrence of hair loss after discontinuation of oral tofacitinib treatment. In the left panel, 8 weeks after discontinuation of treatment is shown and minimal hair loss can be observed. In the right panel, 16 weeks after discontinuation of treatment is shown, and the patient showed almost complete hair loss of the scalp. FIG. 6 shows a biopsy specimen of a scalp from an alopecia areata treated with oral tofacitinib, H & E at baseline (left panel) and 4 weeks after treatment with tofacitinib (right panel) A stained biopsy section of the scalp is shown. FIG. 3 shows hair regeneration during and after luxolitinib treatment. In the upper panel, SALT scores are shown for individual patients during and after discontinuation of ruxolitinib. In the middle panel, the percentage regeneration for individual patients during and after discontinuation of ruxolitinib is shown. In the lower panel, the predicted patient regeneration trajectory (black line) and the actual patient regeneration trajectory (blue line) from the regression model are shown. FIG. 3 shows hair regeneration during and after luxolitinib treatment. In the upper panel, SALT scores are shown for individual patients during and after discontinuation of ruxolitinib. In the middle panel, the percentage regeneration for individual patients during and after discontinuation of ruxolitinib is shown. In the lower panel, the predicted patient regeneration trajectory (black line) and the actual patient regeneration trajectory (blue line) from the regression model are shown. FIG. 2 shows a clinical photograph of a responder AA patient for ruxolitinib, the left panel of each pair is a photograph taken at baseline and the right panel is a photograph obtained at the end of treatment with ruxolitinib It is. Figure 12 shows that biomarkers based on skin gene expression correlate with clinical response, using genes with differential expression between baseline responder samples and healthy control samples, at baseline, 12 weeks of treatment 2 shows a heat map and a clustering dendrogram of samples from patients and samples from healthy controls. FIG. 6 shows that biomarkers based on skin gene expression correlate with clinical response, showing a principal component plot of samples taken from subjects at 12 weeks after treatment and at baseline. It is the figure which showed that the biomarker based on skin gene expression correlates with a clinical response, and has shown the heat map of the ALADIN gene. FIG. 5 shows that biomarkers based on skin gene expression correlate with clinical response, showing a three-dimensional plot of ALADIN signatures. Black, normal subjects; red, AA responder patient at baseline; purple, AA patient after 12 weeks of treatment; yellow, AA nonresponder patient at baseline; blue, AA non-responder after 12 weeks of treatment Responder patient. FIG. 5 shows that biomarkers based on skin gene expression correlate with clinical response, showing ALADIN component signature scores. In the left panel the CTL signature score; in the middle panel the IFN signature score; in the right panel the KRT signature score is shown. * P <0.05, ** p <0.01, *** p <0.001. FIG. 6 shows clinical photographs of selected patients at baseline, end of treatment, and end of observation after the end of treatment. A, D, G, baseline photos. B, E, H, end of treatment. C, F, I, end of observation. FIG. 5 shows an ALADIN signature normalized by a responder procedure. ALADIN component scores from AA patient skin samples were determined at baseline, week 12, and in certain cases at intermediate or post-treatment time points. Blue, responder patient; red, non-responder patient; black, normal control (NC) patient. It is the figure which showed the non-responder patient. A, C, E, baseline photos. B, D, F, end of treatment photo. FIG. 6 shows that gene expression analysis identifies mixed tissue gene signatures, showing unsupervised hierarchical clustering of a cohort of AAP, AT / AU, and unaffected controls (normal) using AAGS. (Blue, underexpressed and red, overexpressed). FIG. 6 shows that gene expression analysis identifies mixed tissue gene signatures, showing a gene common similarity matrix showing gene clusters. The stronger the orange color, the lower the difference in gene expression. Over and under expressed clusters in AA are shown. FIG. 5 shows that gene expression analysis identifies mixed tissue gene signatures, showing a graphical representation of genes in signatures and the statistically enriched functional categories associated therewith. Blue indicates signal transduction pathway; yellow indicates immune / inflammatory pathway; orange indicates HLA; and red indicates cell death pathway. The FDR corrected path with p <0.05 was maintained for this analysis. It is the figure which showed the identification of IKZF1 and DLX4 as MR, and deconvolution of the regulatory factor of the gene expressed in the end organ (skin) from the gene expressed in the infiltrating tissue (immune cell) (decomvo) Shows the overall flow of the pipeline used to End organs (red, AAGS) or invasion based on whether genes measured from complex primary tissue samples (blue green nodes labeled AF) can be mapped to regulators (R) in the skin network Assigned to the object (blue). Only genes mapped to red nodes are considered for MR analysis. Genes mapped to blue nodes are pruned. It is the figure which showed identification of IKZF1 and DLX4 as MR, and is used in order to unfold the regulatory factor of the gene expressed in the end organ (skin) from the gene expressed in the infiltrating tissue (immune cell) Shows the overall flow of the pipeline. The resulting AAGS pruning is overexpressed in AA (cluster 1, node size proportional to fold change) and suppressed (cluster 2), a gene rich in end organs that are mutually exclusive with IGS Providing an expression signature (blue green node), p = 1.77 × 10 −4. It is the figure which showed identification of IKZF1 and DLX4 as MR, and is used in order to unfold the regulatory factor of the gene expressed in the end organ (skin) from the gene expressed in the infiltrating tissue (immune cell) Shows the overall flow of the pipeline. MR analysis was performed on AAGS and IGS in order to solve the convolution of scalp skin control factors to obtain candidate control factors for each signature. True MR (R2) in the skin appears only when AAGS is used and not in IGS. Invasion regulator (R1) is not detected using AAGS. IKZF1 and DLX4 have significant FDR values only when AAGS is used and are not significant when IGS is used (FDR = 1) (left). This analysis establishes IKZF1 and DLX4 (right) as AAGS (blue green square) and MR (yellow square) in the skin. It is the figure which showed identification of IKZF1 and DLX4 as MR, and is used in order to unfold the regulatory factor of the gene expressed in the end organ (skin) from the gene expressed in the infiltrating tissue (immune cell) Shows the overall flow of the pipeline. MR analysis was performed on AAGS and IGS in order to solve the convolution of scalp skin control factors to obtain candidate control factors for each signature. True MR (R2) in the skin appears only when AAGS is used and not in IGS. Invasion regulator (R1) is not detected using AAGS. IKZF1 and DLX4 have significant FDR values only when AAGS is used and are not significant when IGS is used (FDR = 1) (left). This analysis establishes IKZF1 and DLX4 (right) as AAGS (blue green square) and MR (yellow square) in the skin. FIG. 4 shows that exogenous expression of IKZF1 and DLX4 induces a context-independent AA-like gene expression signature, a plasmid vector that expresses IKZF1, DLX4, or an isoform lacking a DNA binding domain. And 2D hierarchical clustering of gene expression measured in huDP and HK transfected with a control expressing RFP. The treatment type and cell type for each experiment are shown at the top of the heat map. Blue indicates reduced expression and red indicates overexpression. 4 shows that exogenous expression of IKZF1 and DLX4 induces context-independent AA-like gene expression signatures, expressed as mean ± SEM, normalized to B-actin. FIG. 6 shows the analysis of IKZF1 and DLX4 mRNA expression in the transfected cells. FIG. 5 shows that exogenous expression of IKZF1 and DLX4 induces context-independent AA-like gene expression signatures, showing a Western blot confirming IKZF1 and DLX4 proteins. FIG. 4 shows that exogenous expression of IKZF1 and DLX4 induces context-independent AA-like gene expression signatures, and measures the specificity of AA-like responses evaluated by differential expression of AAGS after IKZF1 overexpression A GSEA plot is shown. Genes are ranked from left to right on the x-axis from the most differential expression to the least differential expression, and the barcode represents the position of the IKZF1 signature gene. The enrichment score (ES) is shown in the plot and the normalized enrichment score (nES) is displayed at the top. The nES is obtained from the ES at the “tip” of the plot, ie the first maximum ES peak obtained. A p-value is calculated for nES compared against a random null distribution. FIG. 5 shows that exogenous expression of IKZF1 and DLX4 induces context-independent AA-like gene expression signatures, and measures the specificity of AA-like responses assessed by differential expression of AAGS after DLX4 overexpression A GSEA plot is shown. Genes are ranked from left to right on the x-axis from the most differential expression to the least differential expression, and the barcode represents the location of the DLX4 signature gene. The enrichment score (ES) is shown in the plot and the normalized enrichment score (nES) is displayed at the top. The nES is obtained from the ES at the “tip” of the plot, ie the first maximum ES peak obtained. A p-value is calculated for nES compared against a random null distribution. FIG. 3 shows that exogenous expression of IKZF1 and DLX4 induces an increase in NKG2D-dependent PBMC-related cytotoxicity in three cultured cell types. The schematic on the left side of each row describes the tissue introduced into the PBMC for (triplicate) cytotoxicity assays. The color indicates the host source (the matching color indicates the tissue to which the host matches). The middle bar graph shows cytotoxicity values obtained after 6 hours of incubation (all bar heights) or cytotoxicity values observed after 6 hours of addition of human anti-NKG2D monoclonal antibody (grey bars). Represents. NKG2D-dependent cytotoxicity is the difference between these two (white bars). The right bar graph reports changes in NKG2D-dependent cytotoxicity normalized to the RFP control. IKZF1.2B represents cells transfected with the IKZF1δ vector, and IKZF1.3B represents a full-length transcript. The y-axis reports cytotoxicity measured as a percentage of maximum cytotoxicity (total number of cells). All error bars report ± SEM. ** indicates a statistically significant difference from the RFP control with FDR <0.05. (A) It is the figure which showed the data series corresponding to WB215J PBMC and WB215J fibroblast. (B) WB215J PBMC against cultured huDP. (C) WB215J PBMC against cultured HK. The fully reconstituted major regulator module predicts both immune invasion and severity, using exogenous expression data, direct transcription MRT (MR → T), MR It is also possible to estimate T (MR → TF → T) controlled by TF which is T. Any TF (TFB) that is paired with MRs IKZF1 or DLX4 (TFA) and shows a change in expression upon overexpression of TFA is regulated by TFA. Subsequently, any gene (T) in AAGS that is linked to TFB is the secondary T of TFA (to which TFB responds). Since any TFB that does not respond to TFA gene transfer is not controlled by TFA, TFB controls TFA (TFB stable, left) or both are simultaneously controlled by a third TFC (TFB stable, right). FIG. 7 shows that the fully reconstituted major regulator module predicts both immune invasion and severity, and using the approach described above, 78% of AAGS is IKZF1 within one indirect TFB. Or it can be mapped to DLX4. Blue nodes represent AAGS genes that respond to IKZF1 or DLX4 expression, and the size of the nodes is scaled to experimentally observed fold changes (only nodes with at least 25% change are shown) ing). FIG. 4 shows that the fully reconstituted major regulator module predicts both immune invasion and severity and uses the T above to generate a single numerical score for IKZF1 and DLX4 transcriptional activity. , Used to create a classifier for AA severity. The AA sample is then imposed on the search space and evaluated for accuracy (upper figure). The table provides quantification and statistics to separate presentation across regions in the search space (unaffected: NC; mottled AA: AAP; and global alopecia / systemic alopecia: AT / AU). The centroid display can be used to show how the population transitions to the disease state by moving across the trained boundary (bottom figure; non-lesion: AAP-N, And lesion: AAP-L). FIG. 6 shows that the deconvolved control module can be generated for AA, Ps and AD using the same naive framework, and the disease-related gene expression signatures for Ps and AD are expressed differentially. Can be clearly defined by. Comparison of these signatures with AA gene signatures reveals that AA signatures are statistically different from both Ps and AD signatures (Fischer's exact test), and how many of Ps and AD signatures are. There is statistical evidence about such sharing. FIG. 7 illustrates that a deconvolved control module can be generated for AA, Ps and AD using the same naive framework, and converting the above signature into a control module is Compared to the above, it reveals a completely different MR that dominates AD and Ps. Yellow node = AA gene signature; blue node = AD gene signature; indigo green node = Ps gene signature; orange node = AA MR; dark blue node = AD MR; and cyan node = Ps MR It is. A list of the top 5 AD and Ps MRs is provided, ranked by the coverage of the corresponding signature. Also provided is the p-value (IGS p-value) of each MR that has not been unfolded (* indicates a publicly disclosed control factor, and † indicates an MR common to AD and Ps). It is the figure which showed the concentrated path | route in AAGS of FIG. Additional Ingenuity Pathway Analysis shows the enrichment of immune and cytotoxic signaling cascades to both infiltrating populations and end organ processes within AAGS. Genes with differential expression controlled by MR include many membrane-bound proteins, cell death proteins and immune-related proteins. It is the figure which showed the concentrated path | route in AAGS of FIG. Additional Ingenuity Pathway Analysis shows the enrichment of immune and cytotoxic signaling cascades to both infiltrating populations and end organ processes within AAGS. Genes with differential expression controlled by MR include many membrane-bound proteins, cell death proteins and immune-related proteins. FIG. 30 shows the disease gene signatures of AD and Ps of FIG. 29, showing unsupervised hierarchical clustering of lesion and unaffected patient samples using gene expression. Patients are well separated by clinical presentation in psoriasis using associated gene expression signatures. A sample dendrogram is provided here for reference to the heat map provided in FIG. Psoriasis and atopic dermatitis cohorts have gene expression signatures that clearly delineate patients from unaffected controls. FIG. 30 shows the disease gene signatures of AD and Ps of FIG. 29, showing unsupervised hierarchical clustering of lesion and unaffected patient samples using gene expression. Patients are well separated by clinical presentation in atopic dermatitis using associated gene expression signatures. A sample dendrogram is provided here for reference to the heat map provided in FIG. Psoriasis and atopic dermatitis cohorts have gene expression signatures that clearly delineate patients from unaffected controls. FIG. 28 shows the cytotoxicity assay of FIG. 27, where optimization of the PBMC concentration for the cytotoxicity assay identifies a 100: 1 PBMC: target ratio to achieve optimal separation. FIG. 27 illustrates the cytotoxicity assay of FIG. 27, where optimization of the time window for the cytotoxicity assay identifies a time of at least 6 hours to achieve optimal separation. FIG. 6 depicts a list of SNPs for use as biomarkers in connection with the present disclosure. FIG. 6 depicts a list of SNPs for use as biomarkers in connection with the present disclosure. FIG. 6 depicts a list of SNPs for use as biomarkers in connection with the present disclosure. FIG. 6 outlines the design of a clinical study of the treatment of AA with ruxolitinib. It is the figure which described the outline | summary of the state of the research described in FIG. It is the figure which described the outline | summary of the result of the study described in FIG. FIG. 36 depicts the results obtained in connection with the study described in FIG. FIG. 36 depicts the results obtained in connection with the study described in FIG. FIG. 36 depicts the results obtained in connection with the study described in FIG. FIG. 36 depicts the results obtained in connection with the study described in FIG. FIG. 36 depicts the results obtained in connection with the study described in FIG. FIG. 36 depicts the results obtained in connection with the study described in FIG. FIG. 36 depicts the results obtained in connection with the study described in FIG. FIG. 36 depicts the results obtained in connection with the study described in FIG. FIG. 36 depicts the results obtained in connection with the study described in FIG. FIG. 36 depicts the results obtained in connection with the study described in FIG. FIG. 5 outlines the design of a clinical study of the treatment of AA with tofacitinib. FIG. 46 depicts the results obtained in connection with the study described in FIG.

  The subject matter of the present disclosure has the ability to identify patient subpopulations that respond to such treatment with respect to biomarkers that enable improved diagnosis and prognosis of AA and effective treatment for the above-mentioned diseases It includes methods of incorporating biomarkers and methods of incorporating biomarkers that have the ability to track the progress of such treatment.

A. Definitions According to the present disclosure, a “subject” (“subject”) or “patient” is a human or non-human animal. Although the animal subject is preferably a human, the compounds and compositions of the invention are also applied in veterinary medicine, e.g. domesticated species such as dogs, cats, mice and various other pets; Applies to the treatment of livestock species such as horses, sheep, goats, pigs; and wild animals such as non-human primates such as wild or zoos.

  As used herein, the terms “treatment”, “treating” and the like mean obtaining a desired pharmacological and / or physiological effect. This effect may be prophylactic in terms of completely or partially preventing the disease or its symptoms and / or a partial or complete cure for the disease and / or an adverse effect that may result from the disease. It may be therapeutic in terms. “Treatment” as used herein covers any treatment of a disease in a subject or patient, and further: (a) a subject who may be susceptible to a disease but has not yet been diagnosed as having it Preventing the disease from occurring in a person; (b) inhibiting the disease, ie, inhibiting its development; and (c) reducing the disease, ie, reversing the disease; Including.

  A “therapeutically effective amount” or “effective amount” is a compound that, when administered to a mammal or other subject to treat a disease, is sufficient to effect treatment for such disease. Or means the amount of the composition. A “therapeutically effective amount” can vary depending on the compound or composition used, the disease and its severity, and the age, weight, etc., of the subject to be treated.

  The terms “pharmaceutical composition” and “pharmaceutical formulation” as used herein are in a form that allows the biological activity of the active ingredient contained therein to be effective, and Means a composition that does not contain additional ingredients that are unacceptably toxic to the patient to whom the formulation will be administered.

  The term “pharmaceutically acceptable” means, for example, the property of being non-toxic to a subject as used herein with respect to “pharmaceutically acceptable carriers”. The pharmaceutically acceptable ingredient in the pharmaceutical formulation can be an ingredient other than the active ingredient that is non-toxic. Pharmaceutically acceptable carriers can include buffers, excipients, stabilizers, and / or preservatives.

  As used herein, a “JAK inhibitor” is a Jak1 / Jak2 / Jak3 / Tyk2 / STAT1 / STAT2 / STAT3 / STAT4 / STAT5a / STAT5b / STAT6 / OSM / gp130 / LIFR / OSM-Rβ gene, or Interacts with and / or expression of Jak1 / Jak2 / Jak3 / Tyk2 / STAT1 / STAT1 / STAT2 / STAT3 / STAT4 / STAT5a / STAT5b / STAT6 / OSM / gp130 / LIFR / OSM-Rβ protein or polypeptide Means a compound that suppresses This compound may decrease the activity or expression of the protein encoded by Jak1 / Jak2 / Jak3 / Tyk2 / STAT1 / STAT2 / STAT3 / STAT4 / STAT5a / STAT5b / STAT6 / OSM / gp130 / LIFR / OSM-Rβ it can. In certain embodiments, the JAK inhibitor can be a deuterated compound. In certain embodiments, a deuterated compound may be modified by deuteration at one or more sites on the compound.

  JAK inhibitors of the present disclosure may be antibodies (monoclonal, polyclonal, humanized, chimeric or fully human) made against the polypeptides encoded by the corresponding sequences disclosed herein, or binding fragments thereof, etc. It can be a protein. An antibody fragment may be in the form of an antibody other than the full length form, and includes portions or components present within the full length antibody in addition to the engineered antibody fragment. Antibody fragments include single chain Fv (scFv), bispecific antibody, Fv and (Fab ') 2, trispecific antibody, Fc, Fab, CDR1, CDR2, CDR3, combination of CDRs, variable region, quadruple specific Sex antibody, bifunctional hybrid antibody, framework region, constant region, and the like, but are not limited thereto. Antibodies can be obtained commercially, custom generated or synthesized against an antigen of interest according to methods established in the art.

  The JAK inhibitors of the present disclosure can be small molecules that bind to proteins and disrupt their function. Small molecules are a diverse group of synthetic and natural substances that generally have a low molecular weight. Small molecules can be isolated from natural sources (eg, plants, fungi, microorganisms, etc.), obtained commercially and / or available as libraries or collections, or synthesized. Candidate small molecules that modulate proteins can be identified through in silico or high-throughput (HTP) screening of combinatorial libraries. Most conventional pharmaceuticals such as aspirin, penicillin and many chemotherapeutic drugs are small molecules, can be obtained commercially, can be chemically synthesized, or can be obtained from random or combinatorial libraries Can do. In certain embodiments, the agent is a small molecule that binds, interacts with or is associated with the target protein or RNA. Such small molecules can be organic molecules that have the ability to penetrate the cell's lipid bilayer and interact with the target when the target is an intracellular target. Small molecules include, but are not limited to, toxins, chelators, metals and metalloid compounds. Small molecules can be conjugated or conjugated to a targeting agent so as to induce the small molecule specifically in a particular cell.

  In certain embodiments, the JAK inhibitor is ruxolitinib (INCB 018424), tofacitinib (CP690550), tyrophostin AG490 (CAS Number: 133550-30-8), thimerotinib (CYT387), paclitinib (SB1518), valicinib (SB9518), (TG101348), BMS-911543 (CAS Number: 1271022-90-2), Restaurtinib (CEP-701), Fludarabine, Epigallocatechin-3-gallate (EGCG), Peficitinib, ABT494 (CAS Number: 131266-2-60-3) ), AT9283 (CAS Number: 896466-04-9), Desernotinib, Filgotinib, Gand Tinib, INCB39110 (CAS Number: 1334298-90-6), PF 04965842 (CAS Number: 1629022-68-4), R348 (R-932348, CAS Number: 916742-11-5; 1620142-65-5), AZD1480 (CAS Number: 935666-88-9), Serdulatinib, INCB052793 (Incyte, Clinical Trial ID: NCT02265510), NS018 (CAS Number: 1239358-86-1 (free base); 1239358-85-0 (HCl)) , AC 410 (CAS Number: 13141415-84-0 (free base), 136164-15-8-2 (HCl)), CT1578 (SB1578, CAS N mber: 937273-04-6), JTE052 (Japan Tobacco Inc.), PF626263276 (Pfizer), R548 (Rigel), TG02 (SB1317, CAS Number: 937270-47-8), Lambricus reberus extract (AR40, AR40) Pharmaceuticals, LLC.), AR13154 (Arie Pharmaceuticals Inc.), UR67767 (Palapharma S.A.), CS510 (Shenzhen Chipscreen Bioscience Ltd.), VR588 (Vp (D). It is a Ipaforin or a combination thereof.

  In certain embodiments, the JAK inhibitor is specific for the expression of a gene encoding Jak1, Jak2, Jak3, Tyk2, STAT1, STAT2, STAT3, STAT4, STAT5a, STAT5b, STAT6, OSM, gp130, LIFR or OSM-Rβ. Antisense RNA, siRNA, shRNA, microRNA, or a variant or transformation thereof.

B. Alopecia areata biomarkers Embodiments of the present disclosure relate to methods of treating alopecia areata (AA) in a subject. In certain embodiments, a method of treating AA in a subject comprising: disease severity and / or subject before, during and / or after therapeutic intervention in the subject A method is disclosed that includes detecting a biomarker that exhibits a tendency to respond to a person's treatment. In certain embodiments, the biomarker is a gene expression signature. In certain embodiments, the gene expression signature comprises one or more of the following groups of genes: hair keratin (KRT) related genes, cytotoxic T lymphocyte infiltration (CTL) related genes and interferon (IFN) related genes. Contains gene expression information for groups of In certain embodiments, the KRT-related genes include DSG4, HOXC31, KRT31, KRT32, KRT33B, KRT82, PKP1 and PKP2. In certain embodiments, CTL related genes include CD8A, GZMB, ICOS and PRF1. In certain embodiments, the IFN-related genes include CXCL9, CXCL10, CXCL11, STAT1 and MX1.

  In certain embodiments, the gene expression signature is alopecia areata disease activity index (ALADIN). The alopecia areata disease activity index (ALADIN) is a three-dimensional quantitative composite gene expression score for potential use as a biomarker to track disease severity and response to treatment. In certain embodiments, ALADIN is based on gene expression of CTL, IFN, and KRT-related genes, and CTL, IFN, and KRT ALADIN scores are calculated for each sample of the subject. In certain embodiments, a z-score is calculated for each probe set relative to the mean and standard deviation of normal controls. The z-score for each gene may be obtained by averaging the z-scores of probe sets mapped to that gene. In a particular embodiment, the signature score is then calculated with the average of the z-scores for genes belonging to the corresponding signature.

  In certain embodiments, the biomarker is an alopecia areata gene signature (AAGS) comprising one or more genes listed in Table A. In certain embodiments, the biomarker is IKZF1, DLX4, or a combination thereof.

Table A

C. Biomarker Detection Biomarkers used in the disclosed methods can be identified in a biological sample using any method known in the art. Determining the presence of a biomarker, protein or degradation product thereof, the presence of mRNA or pre-mRNA, or the presence of any biomolecule or product exhibiting the expression of a biomarker, or degradation product thereof, is described herein. Or can be practiced for use in any method known in the art.

Nucleic Acid Detection Techniques Any method for qualitatively or quantitatively detecting nucleic acid biomarkers can be used. For example, detection of RNA transcripts can be performed, for example, by Northern blotting. Here, the RNA preparation is run on a denaturing agarose gel and transferred to a suitable support such as activated cellulose, nitrocellulose or glass or nylon membrane. The radiolabeled cDNA or RNA is then hybridized to the preparation, washed and analyzed by autoradiography.

  Detection of RNA transcripts can be further achieved using amplification methods. For example, reverse transcription of mRNA to cDNA followed by polymerase chain reaction (RT-PCR) or using a single enzyme for both steps as described in US Pat. No. 5,322,770 Or R. L. Marshall, et al. It is within the scope of the present disclosure to perform a symmetric gap ligase chain reaction (RT-AGCR) after reverse transcription of mRNA to cDNA as described in, PCR Methods and Applications 4: 80-84 (1994).

  In certain embodiments, quantitative real-time polymerase chain reaction (qRT-PCR) is used to assess biomarker mRNA levels. Biomarker and control mRNA levels can be quantified in affected tissues or cells and adjacent unaffected tissues. In one particular embodiment, the level of one or more biomarkers in a biological sample can be quantified.

  Other known amplification methods that can be utilized herein are described in PNAS USA 87: 1874-1878 (1990) and also described in Nature 350 (No. 6313): 91-92 (1991). The so-called “NASBA” or “3SR” technology; Q-beta amplification as described in European Patent Application Publication (EPA) 4544610; T.A. Walker et al. , Clin. Chem. 42: 9-13 (1996) and strand displacement amplification as described in European Patent Application No. 684315; and target-mediated amplification as described in PCT Publication WO93222461.

  Visualization of in situ hybridization can also be utilized. Here, the radiolabeled antisense RNA probe is hybridized with a thin section of a biopsy sample, washed, cleaved with RNase, and exposed to a sensitive emulsion for autoradiography. The sample can be stained with hematoxylin to verify the histological composition of the sample, and dark field imaging with a suitable light filter shows the developed emulsion. Non-radioactive labels such as digoxigenin can also be used.

  Another method for assessing biomarker expression is to detect the mRNA level of the biomarker by fluorescent in situ hybridization (FISH). FISH is a technique that can directly identify specific regions of DNA or RNA in a cell, and thus allows visual determination of biomarker expression in a tissue sample. The FISH method has the advantage of a more objective scoring system and the presence of a built-in internal control consisting of biomarker gene signals present in all non-neoplastic cells in the same sample. Fluorescence in situ hybridization is a direct in situ technique that is relatively quick and sensitive. The FISH test can also be automated. If it is difficult to determine the expression level of the biomarker by immunohistochemistry alone, immunohistochemistry can be combined with the FISH method.

  Alternatively, mRNA expression can be detected on a DNA array, chip or microarray. Oligonucleotides corresponding to the biomarker (s) are immobilized on the chip and then hybridized with the labeled nucleic acid of a test sample obtained from a subject. A positive hybridization signal is obtained with a sample containing the biomarker transcript. Methods for preparing DNA arrays and their use are well known in the art (eg, US Pat. Nos. 6,618,6796, 6,379,897, 6,664,377, 6,451). , 536, 548, 257, U.S. Patent No. 20030157485 and Schena et al. 1995 Science 20: 467-470; 2000 Drug discovery Today 5: 59-65, which are incorporated herein by reference in their entirety). A serial analysis of gene expression (Serial Analysis of Gene Expression, SAGE) can also be performed (see, eg, US Patent Application No. 20030215858).

  To monitor mRNA levels, for example, mRNA can be extracted from a biological sample to be tested, reverse transcribed, and fluorescently labeled cDNA probes are generated. The microarray can hybridize to the biomarker. The cDNA can then be probed with a labeled cDNA probe, the slide is scanned, and the fluorescence intensity is measured. This intensity correlates with hybridization intensity and expression level.

  Probe types for detection of RNA include cDNA, riboprobes, synthetic oligonucleotides and genomic probes. The type of probe used will generally be dictated by the particular situation, such as riboprobes for in situ hybridization and cDNA for Northern blotting. In certain embodiments, the probe is directed to a nucleotide region that is unique to a particular biomarker RNA. The probe may be as short as required to differentially recognize a particular biomarker mRNA transcript, and may be as short as 15 bases, for example. However, probes of at least 17 bases, at least 18 bases and at least 20 bases can be used. In certain embodiments, primers and probes specifically hybridize under stringent conditions to nucleic acid fragments having a nucleotide sequence corresponding to the target gene. As used herein, the term “stringent conditions” means that hybridization occurs only if there is at least 95% or at least 97% identity between the sequences.

The label form of the probe can be any suitable such as the use of a radioisotope, such as ³² P and ³⁵ S. Labeling with a radioisotope can be accomplished by the use of a suitably labeled base, whether the probe is chemically or biologically synthesized.

Protein detection techniques Methods for the detection of protein biomarkers are well known to those skilled in the art and include mass spectrometry techniques, 1-D or 2-D gel-based analytical systems, chromatography, enzyme-linked immunosorbent assays. immunosorbent assays, ELISAs), radioimmunoassay (RIA), enzyme immunoassay (EIA), Western blotting, immunoprecipitation and immunohistochemistry. These methods use antibodies or antibody equivalents to detect proteins, or use biophysical techniques. Antibody arrays or protein chips can also be used (see, eg, US Patent Application Nos. 2003 / 0013208A1; 2002 / 0155493A1, 2003/0017515 and US Pat. Nos. 6,329,209 and 6,365,418). Reference, which is incorporated herein by reference in its entirety.

ELISA and RIA procedures can be performed to label biomarker standards (using radioisotopes such as ¹²⁵ I or ³⁵ S, or an assayable enzyme such as horseradish peroxidase or alkaline phosphatase), and An unlabeled sample can be contacted with the corresponding antibody, where a second antibody is used to bind to the first antibody and assay for radioactivity or immobilized enzyme (competitive assay). Alternatively, the biomarker in the sample is reacted with the corresponding immobilized antibody, the radioisotope or enzyme-labeled anti-biomarker antibody is reacted with the system, and the radioactivity or enzyme is assayed (ELISA-sandwich assay). Other conventional methods can also be suitably used.

  The above techniques can be performed essentially as a “one-step” or “two-step” assay. A “one-step” assay involves contacting an antigen with an immobilized antibody and contacting the mixture with a labeled antibody without washing. “A two-step assay involves washing the mixture prior to contact with the labeled antibody. Other conventional methods can also be suitably used.

In certain embodiments, the method of measuring biomarker expression comprises the following steps:
Contacting a biological sample, such as blood and / or plasma, with an antibody or variant (eg, fragment) thereof that selectively binds to the biomarker;
Detecting whether the antibody or variant thereof binds to the sample;
Is included. The method can further comprise contacting the sample with a second antibody, eg, a labeled antibody. The method can further include one or more washing steps, eg, to remove one or more reagents.

  It is desirable to be able to immobilize one component of the analytical system on a support so that other components of the system can be contacted with the component and easily removed without tedious labor. It is possible to immobilize the second phase away from the first phase, but usually one phase is sufficient.

  While it is possible to immobilize the enzyme itself on a support, if a solid phase enzyme is required, this will generally be bound to an antibody and supported on supports, models and systems well known in the art. This is best achieved by immobilizing the antibody. Simple polyethylene can provide a suitable support.

  The enzyme that can be used for labeling is not particularly limited, and can be selected from, for example, members of the oxidase group. These catalyze the production of hydrogen peroxide by reaction with its substrate, and glucose oxidase is due to its good stability, availability and low cost, and easy availability of its substrate (glucose) Often used. Oxidase activity can be assayed by measuring the concentration of hydrogen peroxide formed after reaction of the enzyme-labeled antibody with the substrate under controlled conditions well known in the art.

Other techniques can be used to detect biomarkers according to the choice of a practitioner based on this disclosure. One such technique that can be used to detect and quantify biomarker protein levels is Western blotting (Towbin et al., Proc. Nat. Acad. Sci. 76: 4350 (1979)). Cells can be frozen and homogenized in lysis buffer and the lysate subjected to blotting to membranes such as SDS-PAGE and nitrocellulose filters. The antibody (unlabeled) is then contacted with the membrane and a secondary immune reagent such as labeled protein A or anti-immunoglobulin (a suitable label including ¹²⁵ I, horseradish peroxidase and alkaline phosphatase) To assay. Chromatographic detection can also be used. In certain embodiments, immunodetection can be performed using antibodies to biomarkers using an enhanced chemiluminescence system (eg, from PerkinElmer Life Sciences, Boston, Mass.). The membrane can then be stripped and reblotted with a control antibody, such as an anti-actin (A-2066) polyclonal antibody from Sigma (St. Louis, Mo.).

  For example, immunohistochemistry can be used to detect the expression and / or presence of a biomarker in a biopsy sample. A suitable antibody is contacted, for example, with a thin layer of cells, followed by washing to remove unbound antibody, and then contacting with a second labeled antibody. Labeling can be done with fluorescent markers, enzymes such as peroxidase, avidin or radiolabels. The assay can be visually assessed using a microscope and the results can be quantified.

  Other machines or automated imaging systems can also be used to measure biomarker immunostaining results. As used herein, “quantitative” immunohistochemistry scans samples that have undergone immunohistochemistry to identify and quantify the presence of specific biomarkers such as antigens or other proteins. And refers to an automated method of evaluating. The score given to the sample is a numerical representation of the intensity of immunohistochemical staining of the sample and represents the amount of target biomarker present in the sample. As used herein, Optical Density (OD) is a numerical score that represents the intensity of staining. As used herein, semi-quantitative immunohistochemistry refers to the evaluation of immunohistochemistry results by the human eye, in which a trained operator ranks the results numerically (eg, , 1, 2 or 3).

  A variety of automated sample processing, scanning and analysis systems suitable for use in immunohistochemistry are available in the art. Such systems include automatic staining (see, eg, Benchmark system, Ventana Medical Systems, Inc.) and microscopic scanning, computerized image analysis, serial section comparison (to control sample orientation and size variations), digital Report generation, as well as sample archiving and tracking (eg, slides with tissue sections, etc.) can be included. Cellular imaging systems combining a conventional optical microscope and a digital image processing system are commercially available for quantitative analysis of cells and tissues including immunostained samples. See, for example, the CAS-200 system (Becton, Dickinson & Co.).

Antibodies against biomarkers can also be used for imaging purposes, for example to detect the presence of biomarkers in a subject's cells. Suitable labels are radioisotopes, iodine ( ¹²⁵ I, ¹²¹ I), carbon ( ¹⁴ C), sulfur ( ³⁵ S), tritium ( ³ H), indium ( ¹¹² In) and technetium ( ^99m Tc), but fluorescent Includes labels such as fluorescein and rhodamine, and biotin. Immune enzyme interactions can be visualized using different enzymes such as peroxidase, alkaline phosphatase, or different chromogens such as DAB, AEC or Fast Red.

Antibodies and derivatives that can be used include polyclonal or monoclonal antibodies, chimeric, human, humanized, primatized (CDR grafted), veneered or single chain antibodies, phase generating antibodies (eg, phage Display library), as well as functional binding fragments of antibodies. For example, antibody fragments capable of binding to biomarkers, or portions thereof, including but not limited to Fv, Fab, Fab ′ and F (ab ′) ₂ fragments can be used. Such fragments can be produced by enzymatic cleavage or recombinant techniques. For example, papain or pepsin cleavage can generate Fab or F (ab ′) ₂ fragments, respectively. Other proteases with the required substrate specificity can also be used to generate Fab or F (ab ′) ₂ fragments. Antibodies can also be produced in various truncated forms using antibody genes in which one or more stop codons have been introduced upstream of the natural stop site. For example, a chimeric gene encoding the F (ab ′) ₂ heavy chain portion can be designed to include DNA sequences encoding the CH, domain and hinge region of the heavy chain.

  Synthetic and engineered antibodies are described, for example, in Cabilly et al. U.S. Pat. No. 4,816,567, Cabilly et al. EP 0,125,023 B1; Boss et al. U.S. Pat. No. 4,816,397; Boss et al. EP 0,120,694 B1; Neuberger, M .; S. et al. , WO 86/01533; Neuberger, M .; S. et al. , European Patent No. 0,194,276; Winter, US Pat. No. 5,225,539; Winter, European Patent No. 0,239,400 B1; Queen et al. EP 0 451 216 B1; and Padlan, E .; A. et al. , European Patent No. 0519596A1; In addition, Newman, R., et al. et al. , BioTechnology, 10: 1455-1460 (1992) and Ladner et al. , US Pat. No. 4,946,778 and Bird, R., et al. E. et al. , Science, 242: 423-426 (1988)).

  In certain embodiments, agents that specifically bind to polypeptides other than antibodies, such as peptides, are used. Peptides that specifically bind can be identified by any means known in the art, such as a peptide phage display library. In general, an agent capable of detecting a biomarker polypeptide can be used so that the presence of the biomarker is detected and / or quantified. As defined herein, an “agent” refers to a substance capable of identifying or detecting a biomarker in a biological sample (eg, biomarker mRNA, biomarker DNA, biomarker protein Identify or detect). In certain embodiments, the agent is a labeled antibody that specifically binds to the biomarker polypeptide.

  In addition, biomarkers can be analyzed by mass spectrometry, such as MALDI / TOF (time-of-flight), SELDI / TOF, liquid chromatography-mass spectrometry (LC-MS), gas chromatography-mass spectrometry (GC- MS), high performance liquid chromatography-mass spectrometry (HPLC-MS), capillary electrophoresis mass spectrometry, nuclear magnetic resonance spectroscopy, or tandem mass spectrometry (eg, MS / MS, MS / MS / MS, ESI-MS / For example, MS). See, for example, U.S. Patent Application Nos. 20030199001, 20030134304, 20030077616. These are incorporated herein by reference.

  Mass spectrometry is well known in the art and has been used to quantify and / or identify biomolecules such as proteins (eg Li et. Al. (2000) Tibtech 18: 151-160; Rowley et.al. (2000) Methods 20: 383-397; and Kuster and Mann (1998) Curr. Opin. Structural Biol. 8: 393-400;). In addition, mass spectrometry techniques have been developed that allow at least partial de novo sequencing of isolated proteins. Chait et. al. Science 262: 89-92 (1993); Keough et. al. , Proc. Natl. Acad. Sci. USA. 96: 7131-6 (1999); Bergman, EXS 88: 133-44 (2000).

  In certain embodiments, a gas phase ion spectrophotometer is used. In other embodiments, laser desorption / ionization mass spectrometry is used to analyze the sample. Modem laser desorption / ionization mass spectrometry (“LDI-MS”) can be performed in two main variations: matrix-assisted laser desorption / ionization (“MALDI”) mass spectrometry and surface enhancement. Laser desorption / ionization (“SELDI”). In MALDI, an analyte is mixed with a solution containing a matrix and a droplet is placed on the surface of a substrate. The matrix solution is then co-crystallized with the biomolecule. Insert the substrate into the mass spectrometer. Laser energy is directed at the substrate surface. In that case, the substrate surface (it) desorbs and ionizes the biomolecule without significantly fragmenting the biomolecule (them). However, MALDI has limitations as an analytical tool. This does not provide a means for fractionating the sample, and the matrix material can interfere with detection, especially in the case of low molecular weight analytes. See, for example, US Pat. No. 5,118,937 (Hillenkamp et al.) And US Pat. No. 5,045,694 (Beavis & Chait).

  For additional information regarding mass spectrometers, see, for example, Principles of Instrumental Analysis, 3rd edition. Skoog, Saunders College Publishing, Philadelphia, 1985; and Kirk-Othmer Encyclopedia of Chemical Technology, 4th ed. Vol. 15 (John Wiley & Sons, New York 1995), pp. 1071-1094;

  Detection of the presence of a marker or other substance typically includes detection of signal intensity. This in turn can reflect the amount and characteristics of the polypeptide bound to the substrate. For example, in certain embodiments, the signal intensity of peak values from the spectra of a first sample and a second sample can be compared to determine the relative amount of a particular biomarker (eg, visually To computer analysis etc.). Software programs such as the Biomarker Wizard program (Ciphergen Biosystems, Inc., Fremont, Calif.) Can be used to assist in mass spectral analysis. Mass spectrometers and their techniques are well known to those skilled in the art.

Any person skilled in the art will describe any of the components of the mass spectrometer (eg, desorption source, mass analyzer, detector, etc.) and various sample preparations as described in other phrasal components or herein. Or a preparation known in the art. For example, in certain embodiments, the control sample can contain heavy atoms (eg, ¹³ C), thereby mixing the test sample with a known control sample in the same mass spectrometry run. Can do.

  In certain embodiments, a laser time-of-flight (TOF) mass spectrometer is used. In laser desorption mass spectrometry, a substrate having a binding marker is introduced into an inlet system. The marker desorbs and is ionized into the gas phase by a laser from the ionization source. The ions produced are collected by an ion optics assembly and then accelerated in a time-of-flight mass analyzer via a short high voltage field and drifted into a high vacuum chamber. At the far end of the high vacuum chamber, the accelerated ions strike the sensitive detector surface at different times. Since the time of flight is a function of the mass of the ions, the elapsed time between the ion generation and the ion detector impact can be used to identify the presence or absence of a specific mass to charge ratio molecule.

  In certain embodiments, the relative amount of one or more biomarkers present in the first or second sample is determined in part by running the algorithm on a programmable digital computer. The algorithm identifies at least one peak value in the first mass spectrum and the second mass spectrum. The algorithm then compares the signal strength of the peak value of the first mass spectrum with the signal strength of the peak value of the second mass spectrum of the mass spectrum. The relative signal strength is an indication of the amount of biomarker present in the first and second samples. In order to more accurately quantify the amount of biomarker present in the first sample, a standard containing a known amount of biomarker can be analyzed as the second sample. In certain embodiments, the identity of biomarkers in the first and second samples can also be determined.

D. Kits In certain non-limiting embodiments, the present disclosure provides kits for identifying the severity of AA in a patient and for identifying and tracking patient subpopulations that respond to JAK inhibitor treatment. Such kits, in certain embodiments, include means for detecting one or more biomarkers selected from the biomarkers described herein, or combinations thereof. The present disclosure further provides a kit for determining the effectiveness of a therapy for treating AA in a subject.

  In certain embodiments, a kit for treating alopecia areata (AA) in a subject is one or more detection reagents useful for detecting a biomarker indicative of the severity of the subject's disease. And one or more therapeutic reagents useful in the treatment of AA. The subject of the present disclosure is one or more detection reagents useful for detecting biomarkers in a subject that exhibit a tendency to respond to one or more therapeutic reagents useful for treating alopecia (AA), And a kit for treating alopecia areata (AA) in a subject, comprising one or more therapeutic reagents useful for the treatment of AA. In certain embodiments, the kit further comprises one or more probe sets, arrays / microarrays, biomarker specific antibodies and / or beads. In certain embodiments, the kit further comprises instructions. In certain embodiments, the therapeutic reagent may be selected from JAK inhibitors.

  Kit types include, but are not limited to, packaged probes and primer sets (eg, TaqMan probes / primer sets), arrays / microarrays, biomarker specific antibodies and beads. These further contain one or more probes, primers or other detection reagents to detect one or more biomarkers of the present disclosure. Provide a method.

  In certain non-limiting embodiments, the kit comprises a pair of oligonucleotide primers suitable for polymerase chain reaction (PCR) or nucleic acid sequencing to detect one or more identified biomarker (s). Can be included. The pair of primers can include a nucleotide sequence that is complementary to a biomarker described herein and can be of sufficient length to selectively hybridize to the biomarker. Alternatively, complementary nucleotides can selectively hybridize to specific regions 5 'and / or 3' sufficiently close to the biomarker location for PCR and / or sequencing. Multiple biomarker specific primers can be included in the kit to assay multiple biomarkers simultaneously. The kit can also include one or more polymerases, reverse transcriptases and nucleotide bases, which can be further detectably labeled.

  In certain embodiments, a primer can be at least about 10 nucleotides or at least about 15 nucleotides or at least about 20 nucleotides in length and / or up to about 200 nucleotides, or about 150 Up to about 100 nucleotides, or up to about 75 nucleotides, or up to about 50 nucleotides in length.

  In certain embodiments, the oligonucleotide primers can be immobilized on a solid surface or support, such as on a nucleic acid microarray. In that case, the position of each oligonucleotide primer bound to the solid surface or support is known and identifiable.

  In certain non-limiting embodiments, the kit comprises at least one nucleic acid probe suitable for in situ hybridization or fluorescent in situ hybridization to detect the biomarker (s) to be identified. Can do. Such kits generally include one or more oligonucleotide probes having specificity for various biomarkers.

  In certain non-limiting embodiments, the kit can include a primer for detection of a biomarker by primer extension.

In certain non-limiting embodiments, the kit can include at least one antibody for immunodetection of the biomarker (s) to be identified. Both polyclonal and monoclonal antibodies specific for biomarkers can be prepared using conventional immunization techniques, as is generally known to those skilled in the art. The immunodetection reagent of the kit can include a detectable label associated with or linked to a given antibody or antigen itself. Such detectable labels include, for example, chemiluminescent or fluorescent molecules (rhodamine, fluorescein, green fluorescent protein, luciferase, Cy3, Cy5 or ROX), radioactive labels ( ³ H, ³⁵ S, ³² P, ¹⁴ C, ¹³¹ I) or an enzyme (alkaline phosphatase, horseradish peroxidase).

  In certain non-limiting embodiments, the biomarker specific antibody can be provided in association with an individual support, such as a column matrix, an array, or a well of a microtiter plate. Alternatively, the support can be provided as a separate element of the kit.

  In certain non-limiting embodiments, the kit includes one or more primers, probes, microarrays, or antibodies suitable for detecting one or more biomarkers described herein or combinations thereof. be able to.

  In certain non-limiting embodiments, the kit comprises 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 or more of the biomarkers described herein. One or more primers, probes, microarrays, or antibodies suitable for detecting can be included.

  In certain non-limiting embodiments, when the measurement means in the kit uses an array, the set of biomarkers described above is at least 10% or at least 20% for the marker species that appear on the microarray. Or at least 30% or at least 40% or at least 50% or at least 60%, or at least 70%, or at least 80%.

  In certain non-limiting embodiments, the biomarker detection kit includes one or more detection reagents and other components (e.g., buffer, Enzymes, such as DNA polymerase or ligase, chain-extending nucleotides, such as deoxynucleotide triphosphate, and, in the case of Sanger-type DNA sequencing reactions, chain terminating nucleotides, positive control sequences, negative control sequences, etc.). The kit can also include additional components or reagents necessary for detection of the biomarker, such as a secondary antibody for use in Western blotting immunohistochemistry. The kit can further include other prognostic factors, such as one or more other biomarkers or reagents to assess tumor stage.

  The kit can further include a means for comparing the biomarker with a standard and can include instructions for using the kit to detect the biomarker of interest. For example, the instructions indicate that the presence of the biomarkers described herein is for indicating the severity of the patient's AA, or for identifying and tracking patient subpopulations that respond to JAK inhibitor treatment. Can be explained.

  In certain embodiments, the kit can further include a therapeutic reagent. In certain embodiments, the therapeutic reagent can be a JAK inhibitor of the embodiments herein.

E. Reports, programmed computers and systems Tests based on one or more biomarker assays described herein (eg, the severity of an individual AA) or the predicted drug responsiveness of an individual ( For example, response to JAK inhibitor therapy) and / or other information about the study can be referred to herein as a “report”. In some cases, a tangible report can be generated as part of the testing process (this is referred to herein as "reporting" or "providing" a report, A report can be referred to as “creating” or a report “generating”).

  Examples of specific reports include reports in paper (such as printing of computer generated test results) or equivalent formats, and computer readable media (eg, CD, USB flash drive or other removable storage device, computer hard drive) Or a report stored on a computer network server, etc.), but is not limited thereto. Reports, particularly reports stored on computer readable media, can be part of a database. The database may optionally be accessible via the Internet (e.g., having security features that restrict access to the report so that only the patient and the patient's health care provider can view the report) A database of patient records or genetic information stored on a computer network server, which can be a “safe database” that prevents unauthorized individuals from viewing the report). In addition to or instead of generating a specific report, the report can also be displayed on a computer screen (or another electronic device or instrument display).

  The report can include, for example, the severity of an individual's AA, or can only include the presence or absence, level of one or more biomarkers described herein (eg, network server A report on such a computer readable medium includes one or more medical / biological implications such as increased or decreased disease risk for an individual having a specific biomarker or a specific biomarker level Hyperlink (s) to any journal publication or website. Thus, for example, a report can include disease risk or other medical / biological implications (eg, drug responsiveness, suggested prophylactic treatment, etc.), as well as possibly including biomarker information Or can include only biomarker information without including disease risk or other medical / biological significance (eg, from a source external to the report itself, eg, medical personnel, publications) The biomarker information can be used by individuals viewing the report to determine the associated disease risk or other medical / biological significance from the product, website, etc. These are optionally hyperlinked. Etc. to link to the report.)

  Report, for example, individual tested, healthcare professional (eg, doctor, nurse, laboratory technician, genetic counselor, etc.), health care organization, and / or any that is intended to view or possess the report It can be further “transmitted” or “communicated” to other parties or clients (these terms can be used interchangeably herein). The act of “sending” or “transmitting” a report can be performed by any means known in the art based on the format of the report. Further, “sending” or “transmitting” a report includes delivering the report (“pushing”) and / or retrieving the report (“pulling”). can do. For example, the report may be physically delivered by a variety of means, including being physically transferred between parties (eg, for a paper format report), eg, physically delivered from one party to another, or By being transmitted electronically or in signal form (for example via e-mail or the Internet, by facsimile, and / or by any wired or wireless communication method known in the art), for example to a computer network server, etc. It can be transmitted / transmitted, for example, by being retrieved from a stored database.

  In certain exemplary embodiments, the disclosed subject matter provides a computer (or other apparatus / device such as a biomedical device or laboratory apparatus) programmed to perform the methods described herein. . For example, in certain embodiments, the disclosed subject matter receives the identity of one or more biomarkers disclosed herein (ie, input, alone or in combination with other biomarkers). ) And provide disease severity or other outcome (eg, drug responsiveness, etc.) based on the level or identity of the biomarker (s) (ie, as output) ) To provide a programmed computer. Such output (eg, transmission of disease severity, drug responsiveness, etc.) is printed in paper form, for example in the form of a report on a computer readable medium, and / or on a computer screen or other display Can be displayed.

  Certain further embodiments of the disclosed subject matter provide a system for determining the severity of an individual's AA or whether an individual will benefit from JAK inhibitor treatment. Certain exemplary systems include an integrated “loop” where an individual (or their health care professional) requires a determination of the AA severity (or drug response) of such individual, This is done by testing a sample from the individual and the result of this decision is returned to the client. For example, in certain systems, a sample (eg, skin, blood, etc.) is obtained from an individual for testing (the sample can be obtained by the individual or, for example, by a healthcare professional), and the sample is Submitted to a laboratory (or other facility) for (e.g., determining the biomarker (s) disclosed herein, alone or in combination with one or more other biomarkers), And then sending the results of the test to the patient, possibly by first sending the results to an intermediary, such as a health care worker, who provides or otherwise communicates the results to the individual. To form an integrated loop system for determining the severity (or drug response, etc.) of an individual's AA. The part of the system where the results are transmitted (e.g., either between a laboratory, a healthcare worker, and / or an individual) is electronically or by signal transmission (e.g., via e-mail or the Internet, etc.) Can be implemented by telephone or fax, by any other wired or wireless transmission method known in the art, by providing the results on a website or computer network server, which can optionally be a secure database .

  In certain embodiments, the system is controlled by individuals and / or their health care personnel, in which case individuals and / or their health care workers request a test, receive test results, and manage disease. Acting on test results (in some cases) to reduce disease risk, such as by introducing a system.

  Various methods described herein, such as correlating the presence or level of a biomarker with the altered (eg, increased or decreased) severity of AA, etc., are automated methods such as those described herein. By using a computer (or other apparatus / device, such as a biomedical device, laboratory instrument, or other apparatus / device having a computer processor) programmed to perform any of the methods described in can do. For example, computer software (which can be referred to interchangeably herein with a computer program) correlates the presence or absence of a biomarker in an individual with the altered (eg, increased or decreased) severity of AA for that individual. Can be made. Accordingly, certain embodiments of the disclosed subject matter provide a computer (or other apparatus / device) that is programmed to perform any of the methods described herein.

F. Methods of Treatment In certain embodiments, a method of treating alopecia areata (AA) in a subject identifies the subject's AA disease severity by detecting a biomarker indicative of the severity of the disease. And administering to the subject a therapeutic intervention appropriate to the severity of the identified disease.

  In certain embodiments, a method of treating AA in a subject identifies a tendency of a subject with AA to respond to JAK inhibitor treatment by detecting a biomarker that exhibits the trend, and an identified bio If the marker indicates a tendency for the subject to respond to the inhibitor, then administering a JAK inhibitor to the subject.

  In certain embodiments, a method of treating alopecia areata in a subject in need thereof comprises administering a JAK inhibitor to the subject, detecting a biomarker that is responsive to JAK inhibitor treatment, And (1) continuing the administration of the JAK inhibitor (continuing), (2) altering the administration of the JAK inhibitor (altering), or (3) discontinuing the administration of the JAK inhibitor (discontinuing). Optionally tailoring the administration of the JAK inhibitor based on the responsiveness.

  In certain embodiments, the biomarker may be a gene expression signature. In certain embodiments, the gene expression signature includes information on gene expression of one or more of the following gene groups: KRT-related genes; CTL-related genes; and IFN-related genes. In certain embodiments, the KRT-related genes include DSG4, HOXC31, KRT31, KRT32, KRT33B, KRT82, PKP1 and PKP2. In certain embodiments, the CTL-related gene comprises CD8A, GZMB, ICOS and PRF1. In certain embodiments, the IFN-related gene comprises CXCL9, CXCL10, CXCL11, STAT1 and MX1.

  In certain embodiments, when one or more CTL-related genes or one or more IFN-related genes are down-regulated to a predetermined set of gene expression levels, and / or one or more KRT-related genes are predetermined genes When up-regulated to a set of expression levels, the treatment is considered effective and can be continued. In certain embodiments, the majority of CTL-related genes or the majority of IFN-related genes are not down-regulated to a predetermined set of gene expression levels, and / or the majority of KRT-related genes are a set of predetermined gene expression levels. If not up-regulated, the treatment is considered ineffective and may be discontinued or modified, eg, by administering one or more different JAK inhibitors.

  In certain embodiments, the gene expression signature is the circular hair loss disease activity index (ALADIN). In certain embodiments, adapting administration of a JAK inhibitor comprises (1) administering the JAK inhibitor when each of the CTL score, IFN score, and KRT score is reduced compared to the pre-treatment score. Continue, (2) change the administration of the JAK inhibitor if none of the CTL score, IFN score and KRT score are reduced compared to the pre-treatment score, or (3) CTL score, IFN score and KRT Discontinuing administration of the JAK inhibitor if each of the scores is increased compared to the pre-treatment score.

  In certain embodiments, detection of biomarkers that are responsive to JAK inhibitor treatment is performed before the presence of physiological signs of responsiveness to treatment with said JAK inhibitor. In certain embodiments, the detection is performed between 2 weeks and 6 weeks after treatment with the JAK inhibitor. In certain embodiments, the detection is performed 1 week, 2 weeks, 3 weeks, 4 weeks, 5 weeks, 6 weeks, 1 month, 2 months, 3 months after treatment with the JAK inhibitor. After, 4 months, 5 months, 6 months, a combination thereof, or a range between any two of these values.

  In certain embodiments, altering the administration of the JAK inhibitor includes altering dosing intervals, dosages, formulations, or combinations thereof. In certain embodiments, administration of the particular JAK inhibitor may be discontinued and different JAK inhibitors (either different classes of JAK inhibitors or different JAK inhibitors in the same class) administered Good.

  In certain embodiments, the method further comprises establishing a baseline level of a biomarker that is responsive to JAK inhibitor treatment prior to administration of the JAK inhibitor. In certain embodiments, the method further comprises comparing the baseline level with the post-dose level to determine responsiveness to JAK inhibitor treatment before adjusting the administration of the JAK inhibitor.

  In certain embodiments, the detection of a biomarker of the present disclosure is performed on a sample obtained from a subject, and the sample is skin, blood, serum, plasma, urine, saliva, sputum, mucus Selected from the group consisting of semen, amniotic fluid, oral washes and bronchoalveolar lavage. In certain embodiments, the subject is a human. In certain embodiments, the sample is a skin sample. In certain embodiments, the sample is a serum sample.

  In certain embodiments, the detection of the biomarkers disclosed herein is performed via a nucleic acid hybridization assay. In certain embodiments, the detection is performed via microarray analysis. In certain embodiments, the detection is performed via polymerase chain reaction (PCR) or nucleic acid sequencing. In certain embodiments, the biomarker is a protein. In certain embodiments, the presence of the protein is detected using a reagent that specifically binds to the protein. In certain embodiments, the reagent is a monoclonal antibody or antigen-binding fragment thereof, or a polyclonal antibody or antigen-binding fragment thereof. In certain embodiments, the detection is performed via an enzyme linked immunosorbent assay (ELISA), an immunofluorescence assay or a Western blot assay.

  In certain embodiments, the JAK inhibitor is a Jak1 / Jak2 / Jak3 / Tyk2 / STAT1 / STAT2 / STAT3 / STAT4 / STAT5a / STAT5b / STAT6 / OSM / gp130 / LIFR / OSM-Rβ gene or Jak1 / Jak2 / Tyk2 It is a compound that interacts with / STAT1 / STAT2 / STAT3 / STAT4 / STAT5a / STAT5b / STAT6 / OSM / gp130 / LIFR / OSM-Rβ protein. In certain embodiments, the JAK inhibitor may be selected from:

Ruxolitinib (INCB018424):

Tofacitinib (CP690550):

Paclitinib (SB1518):

Balicitinib (LY3009104):

Fedratinib (TG101348):

Decernmotinib:

Restaurtinib (CEP-701):

BMS-915433 (CAS Number: 1271022-90-2), fludarabine, epigallocatechin gallate (epigallocatechin-3-gallate, EGCG), peficinib, ABT 494 (CAS Number: 1310726-60-3), AT 948 (CAS Number: 896466-04-9), filgotinib, gandotinib, INCB 39110 (CAS Number: 1334298-90-6), PF 09658422 (CAS Number: 1622222-68-4), R348 (R-932348, ER Num) 916742-11-5; 1620142-65-5), AZD 1480 (CAS Number: 935666-) 88-9), cerdulatinib, INCB 052793 (Incyte, clinical trial ID: NCT02265510), NS 018 (CAS Number: 1239358-86-1 (free base); 1239358-85-0 (HCl)), AC 410 (CAS Number: 1361645-84-0 (free base); 136164-15-8-2 (HCl)), CT 1578 (SB 1578, CAS Number: 937273-04-6), JTE 052 (Japan Tobacco Inc.) , PF 6263276 (Pfizer), R 548 (Rigel), TG 02 (SB 1317, CAS Number: 937270-47-8), lumbricus rebellus extraction Proceedings, ARN 4079 (Arien Pharmaceuticals, LLC.), AR 13154 (Arie Pharmaceuticals Inc.), UR 67767 (Pala Pharmaceutical SA ur. BioRc. BioRc. BioRc. 04042 (Dynamix Pharmaceuticals / Clevexel), hyperforin, its derivatives, their deuterated variants, their salts, or combinations thereof. In certain embodiments, the detection reagent may be selected from fluorescent reagents, luminescent reagents, dyes, radioisotopes, derivatives thereof, or combinations thereof.

6). EXAMPLES The following examples are presented to provide those of ordinary skill in the art with disclosure and explanation of how to make and use the subject matter of this application. The following examples are not intended to limit the scope of what the inventors regard as the presently disclosed subject matter. In view of the above general description, it is understood that various other embodiments can be implemented.

6.1 Example 1
A. Introduction Alopecia areata (AA) is an autoimmune skin disease in which the hair follicle is the target of immune attack. Patients typically exhibit a rounded or oval patch of hair loss on the scalp, which can spontaneously resolve, persist, or progress to involve the scalp or the entire body. The three major phenotypic variants of the disease are patchy-type AA (AAP), which is often localized to a small oval region on the scalp or in the heel region, and total head hair loss Alopecia totalis (AT) (which involves the entire scalp) and systemic alopecia (alopecia universalis, AU) (which includes the whole body surface area). Currently, there are no drugs approved by the FDA for AA and treatment is often empirical, but typically a wide range of immunosuppressive treatments with observation, intralesional steroids, local immunotherapy or unproven efficacy Accompanied by. More serious forms of the disease, AU and AT, are often resistant to treatment. Furthermore, the prevailing hypothesis among dermatologists and therapists is that the scalp is supported by long-term AU and AT becoming unrecoverable, or by the inverse correlation of response time to treatment duration and responsiveness to treatment. It is to change to a “burned out” state. Despite its high morbidity and the need for effective treatment, the molecular and cellular effector identities of this disease have not been well studied.

  Recent advances in the field have changed the understanding of disease etiology, drug targets, and potential therapeutic solutions. Of particular interest are single nucleotide polymorphisms associated with AA, suggesting that ULBP3 and ULBP6 polymorphisms increase the risk of developing the disease. The ULBP family of genes encodes proteins that function as ligands for NKG2D and, when expressed, mark cells for immune targeting by natural killer cells or NKG2D-expressing CD8 T cells. These data indicate that NKG2D-bearing CD8 in peribulbar infiltrates in skin sections of lesion scalp biopsy specimens of AA patients and in affected skin and skin draining lymph nodes from the C3H / HeJ mouse model of spontaneous AA. It led to recognition of T cells. Adoptive transfer of this cell population from C3H / HeJ mice with alopecia to unaffected C3H / HeJ mice results in induction of alopecia and demonstrates the central role of these effector cells in the mouse AA model did.

  We have previously identified prominent interferon (IFN) signatures and common gamma chain cytokine (γc) signatures. Both of these were postulated to contribute to AA pathogenesis. Based on these findings, therapeutic strategies based on the inhibition of important members of the signaling molecule, Janus kinase (JAK) family, are effective in treating AA in mouse models of disease and a small set of human patients It was found that Gene expression profiling played an important role in the selection of small molecule JAK inhibitors for AA. And expanded efforts in this regard, including different AA phenotypes, have the potential to provide new therapeutic solutions as well as further insights into pathogenic mechanisms.

  In this study, a total of 96 patients with a series of AA phenotypes and over 120 samples taken from normal control patients were profiled. Patient samples were collected from the National Registry of Alopecia areata after phenotypic classification by a dermatologist specializing in hair disorders. The skin biopsy samples were then examined using microarray-based gene expression analysis to identify AA-specific gene expression signatures. Despite the general notion that AT / AU is a “burned out” form of disease or cannot be recovered, gene expression analysis found a significant amount of immune activity in AT / AU samples Thus, it has been shown that a treatment that destroys this immune activity may be useful for therapeutic purposes. Furthermore, an alopecia areata disease severity index (ALADIN) was created based on the data. This can be used to effectively distinguish AT / AU, AAP, and NC samples from each other and to track disease activity in patients undergoing conventional or experimental treatment. It was a gene expression metric.

B. Results AA gene signatures Gene expression profiling was performed on 122 samples from 96 patients consisting of a 63 patient discovery data set and a 33 patient external validation data set (for a more complete explanation). , See method section). Microarray-based gene expression analysis was performed on a discovery data set consisting of 20 AAP, 20 AT / AU, and 23 normal control scalp skin biopsy specimens. Differentially expressed genes were identified based on a comparison of AA sample versus normal controls. To ensure the robustness of the data from this initial sample set, external validation was performed using additional 8 AAP, 12 AT / AU, and 13 normal control scalp skin biopsy specimens as a validation set. From this analysis set, a disease-specific gene expression profile is generated based on genes with differential expression selected with fold change (FC)> 1.5 and false discovery rate (FDR) <0.05. It was done. The AA-specific disease signature consisted of 1083 Affymetrix probes showing increased expression and 919 Affymetrix probes showing reduced expression in AA.

  Of note, genes associated with cell-mediated cytotoxicity, including PRF1 and several granzymes, as well as immune cell transport chemokine genes are among the top-level genes listed as showing increased expression. On the other hand, hair keratin related genes such as DSMG4, FGF18 and GPRC5D and developmental genes were among those genes that showed reduced expression. The pattern of gene expression distinguished the phenotype groups from each other. Here, the normal control and AT / AU samples show the greatest discrepancy (FIG. 1A). Plotting the sample on a terrain expression map revealed three clusters corresponding to healthy controls, AAP patients, and AT / AU patients. These patient groups descended along a nearly linear path through the terrain map (FIG. 1B).

  Based on its location within this terrain, a single score was generated that assesses the relative risk of any given sample that is AAP or AT / AU. This score is in turn based on a consensus of genes that have differential expression between AA and healthy controls (see methods). The resulting score is between 0 and 10, where 10 represents the risk of maximum severity (AT / AU) and 0 represents the minimum risk (healthy control). The AAP sample was in the middle range of these two extremes (score range 2-6). Both AAP and AT / AU cohorts had statistically separable mean scores compared to healthy controls (box-and-whisker plot in FIG. 1B). Genes that were differentially expressed from the discovery data set were able to distinguish AA samples from normal samples by hierarchical clustering in the validation set (FIG. 6). These data suggest that AA pathology can be expressed at the molecular gene expression level and that the AAP sample exhibits an AA-specific signature that is intermediate between AT / AU and normal controls.

AT / AU skin samples are immunologically active A linear representation of the molecular classification between control, AAP and AT / AU in a general gene expression analysis combined with the presence of immune-related genes in disease signatures It was suspected whether the AU sample was immunologically active. Since the AT / AU sample appeared to show a more severe AA-specific signature than that of AAP based on both the differential expression level and the number of differentially expressed genes, the AT / AU gene expression profile, and The gene expression profile (that) for AAP compared to healthy controls was examined separately. AT / AU-specific disease signatures based on FC> 1.5 and FDR <0.05 were composed of 2239 genes with increased expression and 1643 genes with decreased expression. AAP-specific disease signatures based on similar thresholds show genes with fewer expression differences, where only 376 Affymetrix probes have increased expression and 537 Affymetrix probes have reduced expression Had. Comparison of the AT / AU-specific gene list and the AAP-specific gene list showed duplication of AAP-specific genes in the two lists. Here, AAP-specific genes were hardly included in the AT / AU-specific gene list (FIG. 2A). In addition to the previous data, these data indicate that AT / AU is more complex and more complex than the more localized AAP form of the disease, in contrast to the hypothesis that AT / AU is “burned out” It shows that it is serious.

  Focusing on the more robust expression of immune-related genes, we attempted to support the more severe gene expression profiles found in AT / AU samples using another method. To determine the extent to which T cell infiltration was observed in AT / AU samples, which are the most abundant and most functionally relevant invasive immune cells in AA compared to AAP samples, pan-T cells Immunohistochemical staining of marker CD3 was performed on AT / AU, AAP, and NC samples. The AT / AU sample showed a significant increase in relative immune infiltration compared to that of NC samples (FIG. 2B). Then, using a perispheric / peri follicular invasion histopathological scoring system (HPS), an additional trend of invasion was further observed compared to AIP samples (FIG. 2C). These data indicate that AT / AU samples show high and sustained amounts of immune activity and inflammation.

  Pathway analysis was performed on signatures up-regulated with either AAP or AT / AU samples. Interestingly, in both AAP and AU / AT, including “graft versus host disease”, “type I diabetes mellitus”, “allograft rejection”, “cell adhesion molecules” and “antigen processing and presentation” A shared set of up-regulated pathways (FIGS. 2D-2E) is composed of antigen presenting genes and supports the pathogenic theme of loss of immune privilege of hair follicle microenvironment and immune activation in AA. Interestingly, the “chemokine signaling pathway” was also found to be significantly up-regulated, and as suggested for other autoimmune skin diseases, these intercellular transport molecules were used for therapeutic purposes. Increased the possibility of targeting. These results indicate that the majority of active immune pathways in AA are the same in the milder forms of the disease as well as in the more severe forms of the disease.

Paired lesional and nonlesional samples are similar in gene expression profile AA-specific gene signatures are present after the onset of symptoms or rather unaffected subjects It is unclear whether it exists as part of a global signature that can be used to distinguish AA subjects. To test whether non-lesional skin from AA patients is significantly different from normal patient's skin, an additional 18 samples of non-lesional skin from AA patients (AAP-L) served as training set irregularities Analysis on the affected skin (AAP-LS). And 8 additional AAP-NL samples with corresponding AAP-LS were used for the validation set. Based on FC> 1.5 and p-value <0.05, genes with differential expression between AAP-LS skin and AAP-NL were determined. Here, 27 genes show increased expression of AAP-LS and 143 genes show reduced expression of AAP-LS. Examination of the expression level of this set of genes showed that AAP-NL shows a profile that is intermediate between AAP-LS and NC (FIG. 3A).

  In order to more clearly visualize the differences between these three populations detectable by gene expression, a principal component (PC) analysis was performed. This analysis attempts to identify the smallest dimension in highly complex data that maximizes sample separation based on gene expression. The end result of this analysis is that samples with similar gene expression profiles cluster closely with each other and samples with dissimilar profiles cluster separately from each other. The non-lesion sample (AAP-NL) showed very variable dissimilarities from the patient-matched lesion sample (AAP-LS, FIG. 3B) based on the first component of the PC analysis (AAP-LS, FIG. 3B). Arranging all samples for the first two principal components was consistent with this data. Here, AAP-LS and NC samples showed heterogeneous clustering, and AAP-NL showed a profile intermediate between AAP-LS and NC (FIG. 3C). A plot of the first component of the PC analysis of 8 AAP-L / AAP-NL pairs from the validation data set showed the same highly variable dissimilarity across the observed pairs in the discovery data set. Genes that were differentially expressed between non-lesion and non-lesion samples were analyzed for functional annotation. And it was found that the most common genes present in non-lesion samples (but not in lesion samples) were hair-related keratins and a handful of inflammatory response genes (FIG. 3D). Genes associated with immune response and invasion, including CCL5 / 13, PRF1, GZMB / K, ITGAM, and CD209, were missing from non-lesional samples. These results indicate that a non-lesion sample biopsy from an AA patient does not completely resemble the lesion area but does not resemble a healthy control scalp. These samples are different from normal, unaffected samples, but exist in an intermediate state that has not yet elicited a complete autoimmune response as seen in lesion samples.

Invasion gene expression signatures correlate with the AA phenotype The presence of significant infiltrates in both the AAP and AT / AU patient cohorts, and the presence of immune-related marker genes in the gene expression array cohort It was suspected whether it could be detected directly. The ability to detect an infiltrating population will prove beneficial for the pathology and characterization of AA. In order to identify invasive immune tissue, a unique gene expression signature was employed that defined each of several infiltrating immune cells and was used as an immune gene signature (IGS). This work provided comprehensive and mutually exclusive genetic markers for several immunotypes including but not limited to B cells, T cells, macrophages, natural killer and mast cells. Numerical relative measurements of the relative invasion of each of these tissue types were constructed as a function of the corresponding IGS expression (FIG. 4A). Using this metric, we were able to quantify the relative infiltration of each immune tissue for each patient and test them for correlation with AA initiation and severity (FIGS. 7A- 7B). Of the infiltrates tested, only CD8 T cell and natural killer cell rankings had the ability to separate NCs from AAP or AT / AU. Ranking by CD8 activity resulted in a dose-dependent separation between the three clinical symptoms and significantly separated the three populations (the hash represents the median value of each cohort). Although NK-specific markers did not reflect the ability of CD8 T cell-specific markers, it indicates that the correlation may not be a result of NK invasion or shared NK / CD8 T cell genes.

Using IGS metrics, we also have total invasion signals that contaminate AAP samples (Figure 4B, left pie) and AT / AU samples (Figure 4B, right pie). Estimated. The overall estimated change in invasion of each immune tissue type is also shown (FIG. 4B, chart). From gene expression data, an estimated infiltrate contamination of 0.8-1.4% was observed and correlated with increased clinical severity of AA. Consistently, CD8 ⁺ infiltrates consisted of more than 65% of the total invasive load only in samples from AAP or AT / AU patients. An absolute change in each immune tissue infiltration across the three presentations is also shown (FIG. 4C), indicating that only CD8 + infiltration changes significantly across the three populations. These results indicate that the most significant existing type associated with AA is non-NK, CD8 ⁺ cells, although there is evidence based on some expression for multiple invasive tissue types. In addition, we were able to detect elevated levels of markers associated with macrophages, total CD4 ⁺ T cells, CD4 ⁺ T cell subsets, NK cells and B cells, which are CD8 T cell compartments. A small fraction was shown compared to the fraction. Overall, the contamination within each sample is relatively small and the total population of immune tissues tested does not exceed about 3% of the total signal.

In addition, IGS scores were used to estimate the relative Th1 and Th2 fractions detected in patient samples (FIG. 4D). For each patient (AA or unaffected control), the Th load within the sample biopsy is expressed as a ratio of Th1: Th2 signal, and AA patient samples show a shift to a higher Th1 ratio compared to normal controls. It was observed. The Th1: Th2 rank shift associated with AA presentation is statistically significant by Mann-Whitney U test (p = 1.02 × 10 ⁻⁴ ), and overall, skin from AA patients is Shows that there are AA patients that contain elevated levels of Th1 signatures associated with Th2 signatures but have both Th2 and Th2 signatures compared to unaffected patients.

The ALADIN score is parallel to the disease phenotype We sought to create a metric that identified the most prominent features of the AA disease signature that would allow a quantitative assessment of the disease state. Weighted gene co-expression analysis (WGCNA) of genes with differential expression between AA and healthy controls revealed 20 clusters of co-expressed genes (FIG. 5A). These gene sets represent co-expressed modules, indicating the potential for co-regulation, shared biological functions and / or shared pathways. For each of these modules, we define a color-coded eigengenes or metagenes using the first principal component of the gene expression signature derived from the genes in each module. can do. The gene set enrichment analysis (GSEA) of these modules with a ranked list of genes differentially expressed between the AA and NC cohorts, as well as module metagenes and disease phenotypes The association test between the green and brown modules is most significantly associated with the disease phenotype, and these modules (FIGS. 5B, 8A and 8B) (Green) and structural keratin (brown) were revealed. Green module pathway enrichment analysis revealed several gene pathways associated with autoimmune responses (FIG. 5C). This included genes such as CD8, CD4, MICB, CCL4 / 5, CCR7, and ICOS. Both perforin and granzyme B were detected, as well as genes previously involved by the GWAS meta-analysis, including ICOS, IRF1 and CIITA.

  An original scoring system, the Area Disease Activity Index, ALADIN, was developed. This was a three-dimensional quantitative complex gene expression score because of its potential use as a biomarker to track disease severity and response to treatment. The metric evaluates patients along a combination of cytotoxic T lymphocyte infiltration (CTL), IFN-related markers (IFN), and hair keratin panel (KRT). Interestingly, the CTL signature contains two genes, CD8A and PRF1. These constitute the CD8 T cell signature described above (FIG. 4). Examination of the components of the green module revealed the presence of genes in both ALADIN CTL and IFN signatures. And the Brown signature contained the genes that make up the ALADIN KRT signature. To test the robustness of ALADIN against a new data set, patient CTL, IFN and KRT scores were determined (Figure 9). A three-dimensional plot for the combined discovery and validation data set of 96 AT / AU, AAP, and NC samples showed that the AT / AU sample was clustered furthest from the NC sample. Here, the AAP sample was placed in an intermediate position between both of these sets (FIG. 5D). Subsequent GSEA showed a statistically significant enrichment of the original ALADIN gene set in AA samples compared to normal controls in both AAP and AT / AU cohorts (FIG. 8C). These data indicate that the ALADIN score can distinguish AA forms of different severity and invites the use of this metric in clinical trials.

  In addition, we evaluated whether the duration of the disease affected the ALADIN score. Skin samples from AT / AU patients who have been ill for more than 5 years showed a statistically significant decrease in IFN and CTL scores when compared to samples from AT / AU patients with shorter durations (FIG. 5E). . This relationship was not seen between the long-term and short-term AAP samples (Figure 10). These data show that the ALADIN score can distinguish AA forms of different severity and, moreover, inflammation and immune infiltration scores decrease over time among the more severe forms of AA Yes.

C. Discussion A microarray-based whole-genome gene expression assay was utilized to provide basic insights into AA biology. The work here involves the use of more than 120 scalp biopsy specimens from patients with AA and healthy controls. We use this method for the first time to identify some important features of pathogenesis.

  First, AT / AU exhibits a relatively high level of immune activity compared to normal controls and AAP samples. The notion that patients with AT / AU cannot be treated effectively is the historical difficulty of treating these patients with previously available topical and oral medications and the considerable unsettledness in skin biopsy specimens of critically ill patients. It appears to be due to the difficulty in identifying rudimentary hair. However, this data challenges this idea by providing evidence of sustained immunological activity in AT / AU samples that is equal (if not greater) to that seen in AAP. This immune activity in patients with AT / AU is rare enough for spontaneous resolution of AT / AU disease, but in combination with case reports, is sufficiently strong immunity to target the pathways required to maintain the immune response It means that an inhibitor or treatment may be effective for these types of patients. In fact, recent mechanistic data support the role of the Janus kinase-mediated pathway in AA, and some case reports show that small molecule JAK inhibitors are drugs for AA, even in severe or widespread diseases. It confirms that it seems to be a promising class.

  Second, the molecular definition of AA supports a prominent role for CD8 T cells in the pathogenesis of human diseases. When comparing NC, AAP and AT / AU samples, a relationship such as a dose response can be seen, and pericyte / follicular T cell trends can also be observed. Previous studies have shown that CD8 T cells are necessary and sufficient in a mouse model of AA. It suggested a role for CD8 T cells due to the expression of KG2D and the association found between AA and NKG2DL in AA pathogenesis. The data not only further support the role of CD8 T cells in disease pathogenesis, but also elicit a correlation between the level of CD8 T cell involvement and the severity / phenotype of the disease.

  Third, the relative similarity between the non-lesioned AAP skin sample and the lesioned AAP skin sample indicates that the unaffected skin in AA patients compared to the relationship observed between the lesioned AAP and normal skin samples. , Suggesting that the skin of AA patients shares at least some of the factors that direct hair loss to develop. However, due to the clinical lack of disease, non-lesional skin samples do not have a complete constellation of the factors necessary for the autoimmune destruction of hair follicles. For signs of clinical disease, it appears that some immune dysregulation to resistance collapse must be overcome. Here, alopecia is observed only when all of these situations occur. Further examination of the difference between unaffected and affected skin samples from AA patients will likely reveal new treatment or prevention strategies in predisposed individuals, and ultimately Can be beneficial for various stimulation triggers.

  The main part of this work is to establish a molecular definition of the disease process in the skin and to examine the signatures corresponding to protein mediators or cellular parties. These data serve as an abundant resource for researchers seeking pathogenic disease mechanisms and therapeutic targets in AA.

D. Materials and Methods Human Patient Demographics Two independent data sets were collected from four National Alopecia Area Foundation (NAAF) registration sites. The discovery data set consisted of 81 samples from 63 patients (20 AAPs, 20 AT / AUs, and 23 normal controls. 18 AAPs also had surrounding and non-lesioned AAP lesions. Contributed biopsy of non-lesioned skin to allow a pairwise comparison of gene expression between. The validation data set consisted of 41 samples from 33 patients. (8 AAP, 12 AT / AU, and 23 normal controls. 8 of the AAP patients also allowed a pairwise comparison of gene expression between AAP lesioned and non-lesioned samples. Contributed biopsy of non-lesional skin.)

Human Tissue Sampling and Processing Skin punch biopsy specimens were fixed in PAXgene Tissue Containers and shipped to Columbia University overnight. The sample was bisected. Half of the sample was processed with the PAXgene tissue miRNA kit to extract RNA, and the other half was embedded in paraffin. Library preparation (prep) was performed for microarray analysis using Ovation RNA Amplification System V2 and Biotin Encore kit (NuGen Technologies, Inc., San Carlos, Calif.). Samples were then hybridized to the human genome U133 plus 2.0 chip (Affymetrix, Santa Clara, Calif.) And scanned on Columbia University Pathology Core or Yale Center for Genome Analysis (YCGA).

Analysis Package Quality control of the microarray was performed using the affy Analysis QC package from http://arrayanalysis.org/. It was a batch effect correction. Differential expression in these studies was defined with an absolute fold change threshold of 1.5. Significant difference threshold value 0.05 was corrected by Benjamini-Hochberg. Clustering and principal component analysis were performed using modules provided in the Bioconductor R package. A network image was generated with Cytoscope.

Microarray Pretreatment and Quality Control Microarray pretreatment was performed using a BioConductor at R. Pre-processing of the two data sets, the discovery data set (63 samples) and the validation data set (33 samples) was performed separately using the same pipeline. Quality control is available at http://arrayanalysis.com. This was performed using the affanalysis QC package from org /. The discovery and validation data sets were normalized separately using GCRMA and MAS5. The Affymetrix HGU-133Plus2 array houses the 54675 probe set (PSID). Filtering was performed. Thereby, PSIDs that were on the X or Y chromosome, PSIDs that were Affymetrix control probe sets, or PSIDs without the Gene Symbol annotation were removed from all arrays for further downstream analysis. For 3D plots of ALADIN scores, all 96 samples from both data sets were combined prior to GCRMA normalization and batch effect correction.

Sample Filtering Batch Modification To perform analysis on 63 AA lesions (both AT / AU and AAP) and NC samples in a discovery data set, called presence on at least one 63 array that yields a 36954 PSID PSIDs were further filtered to remove missing PSIDs. Correction for batch effects was made using the implementation of the function ComBat available in the sva package with gender and AA groups (AT / AU, AAP, and normal) used as covariates. The validation set did not require batch correction. Paired lesion / non-lesion microarrays were processed together in the same microarray batch along with normal controls. The discovery set for paired lesion / non-lesion analysis consisted of 18 lesion / non-lesion AAP pairs and 23 controls. The validation set had 8 lesion / non-lesion AAP pairs and 13 NC samples.

  To examine the relationship between paired lesion and non-lesion samples, PSID was focused on the average expression level of normal samples within each batch. The validation set did not require batch modification.

Differential expression analysis
Differential analysis was performed on the batch-corrected discovery data set using a linear model implemented in the limma package at Bioconductor. Two sample comparisons were performed separately to identify PSIDs with differential expression in AA patient versus normal control, AAP patient versus normal control, and AT / AU patient versus normal control. Gender was treated as a fixed factor. A pairwise comparison was performed on AAP-LS / AAP-NL samples that treated gender as a fixed effect.

  In order to reduce the number of false discoveries retained in gene expression signatures, we subsampled the discovery data set because log2 (fold-change) did not exceed 1 for most PSIDs. Samples from one batch at a time were leaving out and only PSIDs identified as differentially expressed in all datasets and all subsamplings were retained.

Principal component analysis Principal component analysis was performed on all 36954 PSIDs used to perform differential expression analysis. The probability density of the first two principal components was estimated for each group (AT / AU, AAP, NC), assuming a bivariate distribution. Principal component analysis was performed. The 41 AAP-LS, AAP-NL, and normal control samples in the discovery set used 170 PSIDs identified as differentially expressed.

Histopathological staining Immunohistochemistry was performed using the Bond Polymer Refine Red Detection (Leica Biosystems, Buffalo Grove, IL) protocol using a clone LN10 anti-CD3 primary antibody. Peribulbar / peri follicular histopathology scoring system (HPS) is a representative example shown in FIG. 2C with 0-3 scale (0-no immune infiltration; 1-mild; 2-medium; 3-severe) It was performed using. Using the Kruskal-Wallis test, the null hypothesis that the average rank of the CD3 score was the same at the 0.05 significance level in all groups was rejected (H = 16.51, df = 2, p = 0). .00026). The Wilcoxon rank sum test was performed for all pairwise comparisons and adjusted for multiple comparisons using Bonferroni correction. Significant differences were observed between the AU / AT group and normal group (p = 4e-04) and the AAP group and normal group (p = 0-4) (distribution = R using asymptotic) Wilcox_test from coin package).

Generating a linear Euclidean classifier for AA severity A principal component analysis and topographic mapping of AA disease signatures are performed in the expression space defined by the first two principal components (PC), NC, AAP, And revealed near linear dependence between AT / AU patients. The expression terrain map was generated with the MeV software suite using Euclidean distance as a metric and 10 nearest neighbors as clustering parameters. In order to translate this into a more intuitive numerical score that predicts severity, we generated a list of genes that contribute significantly to PC1 and PC2. This was done by rank-sorting the gene's weighted contribution to each PC and selecting the set of genes before the inflection point of the weight distribution. The expression vectors for these genes were then z-score transformed and rank-normalized to generate non-zero, statistically comparable expression values. For each patient, all genes in each normalized vector were used to construct a centroid value in the appropriate PC vector. Each centroid value then corresponds to a grid base point defined as PC1xPC2 for each patient. Next, a linear projection was constructed between {PCminxPC2min} and {PClmaxxPC2max} and each patient was mapped to this line. The vector was then normalized to link values between 0 and 10. The score breakpoint (NC 0-2, AAP 2-6, AT / AU 6-10) for each cohort is the score that maximizes the odds ratio of NC and AAP and AAP and AT / AU within each score range Obtained by performing a sliding window analysis to identify values.

Generation of consensus immune gene signatures Unique signature genes were employed for each infiltrating population. Independently test each group of mutually exclusive genes for cosegregation and discriminatory power before consolidating them into a consensus score for each individual patient to generate relative classifiers that rank infiltration did. Integration was performed in the following steps: z-score to transform all gene vectors to obtain scale-comparable expression values; these vectors were used to obtain ordered non-negative values for each gene signal. Rank transforming; and finally averaging ranks to create a rank order consensus value for each infiltrating tissue. For estimation of total invasive load per sample, the consensus z-score was converted back to the expression space of each individual gene and normalized to the consensus of housekeeping genes with the smallest coefficient of variation across the population. .
The following table shows the signature for each cell type.

Unsupervised machine learning, weighted gene co-expression analysis Weighted Gene Co-expression Analysis (WGCNA) is a ComBat-regulated expression set that exceeds the median of all PSID variations. For PSID. Adjacency between PSIDs was defined as the Pearson correlation of the expression profile across the sample raised to a soft thresholded output set to 10. The obtained adjacency matrix was converted into a topological overlap matrix (TOM) from which a difference matrix was calculated. Hierarchical clustering was performed on dissimilar matrices, and modules were identified from the resulting branches of the dendrogram. The co-expression heat map and dendrogram in FIG. 5 were generated from stratified sampling of 1/3 of the PSID assigned to 20 modules, using TOM as an adjacency criterion. To test the association between module-specific genes (eigengenes) and categorical traits, disease phenotype, sex, and origination of NAAF sites, the Kruskal-Wallis test was performed. Pearson's correlation coefficient and p-value were estimated between each module specific gene and subject age. Gene set enrichment analysis (GSEA) was used to further test for over-expression of genes that were over / under-expressed in AA for normal controls in different WGCNA modules. The normalized enrichment score (NES) reflects the degree to which gene sets are over-represented at the top or bottom of the ranked gene list, taking into account differences in module size. A pre-ranked gene list was created with t statistics for ranking.

Calculation of ALADIN score CTL, IFN and KRT ALADIN scores were calculated for each sample. Briefly, a z-score is calculated for each PSID compared to the mean and standard deviation of normal controls. The z-score for each gene is obtained by averaging the z-score of PSID mapped to that gene. The signature score is then the calculated average of the z scores for the genes belonging to the corresponding signature.

6.2 Example 2
A. Introduction AA is a T cell-mediated autoimmune disease that is characterized phenotypically by alopecia and histologically by infiltrating T cells surrounding the hair follicle. Migration of total T cells (but not B cells or serum) can cause disease in human xenograft models as well as in C3H / HeJ mice, a mouse strain that develops spontaneous AA much similar to human AA. . Broadly acting intralesional steroids are the most commonly used therapy for AA and have had varying success. Advances in the development of rationally targeted and effective therapies have been limited by a lack of mechanistic understanding of the underlying major T cell inflammatory pathway in AA.

  A cytotoxic subset of CD8 + NKG2D + T cells was identified within the infiltrate surrounding human AA hair follicles. A “risk signal” ULBP3 and MICA, two NKG2D ligands (NKG2DL), and associated upregulation in the hair follicle itself of the two NKG2D ligands were also identified. Its importance in disease pathogenesis has also been suggested by genome-wide association studies.

B. Results To determine the contribution of CD8 + KG2D + T cells to AA pathogenesis, we spontaneously developed alopecia and recapitulates many pathological features of human AA (C3H / HeJ). A mouse model was used. In a lesion skin biopsy of alopecia mice, CD8 + KG2D + T cells infiltrate the epithelial layer of the hair follicle. This overexpresses KG2DL, H60 and Rae-1 similar to those observed in human AA skin biopsies (FIGS. 11A-11B). Flow cytometric analysis of the CD45 + leukocyte population in the skin shows that CD8 + KG2D + T cells in the skin of affected C3H / HeJ mice combined with increased skin lymphadenopathy and total cellularity compared to disease free C3H / HeJ mice. A significant increase in numbers was revealed (FIGS. 11C-11D). There were a much smaller number of other cell types, including CD4 + T cells 4 and mast cells.

  The immunophenotype of skin infiltrating CD8 + T cells in mice with AA was similar to that of the CD8 + KG2D + population found in cutaneous lymph nodes: CD8αβ + effector memory T cells (TEM, CD8hiCD44hiCD62LlowCD103 +) are CD49b and NKG2A, NKG2C and NKG2E. Retains multiple natural killer (NK) immunoreceptors, including (FIG. 11E). These CD8 + TEM cells expressed high levels of IFN-γ and exhibited NKG2D-dependent cytotoxicity against syngeneic dermal sheath target cells grown ex vivo (FIG. 11F). Gene expression analysis of CD8 + KG2D + T cells isolated from alopecia C3H / HeJ lymph node cells using RNA-seq showed a transcriptional profile characteristic of effector cytotoxic T lymphocytes (CTL) and several Additional NK specific transcripts were identified.

  We next evaluated the need for these CD8 + TEM cells in the pathogenesis of the disease. Metastasis of cytotoxic CD8 + KG2D + cells or whole lymph node cells from diseased mice induced AA in all 5 healthy C3H / HeJ recipients by 14 weeks post-transplantation, while lymphs depleted of NKG2D + cells The nodal cell population failed to metastasize the disease (FIG. 11G). Thus, CD8 + KG2D + T cells are the dominant cell type in dermal infiltrates and are necessary and sufficient for T cell mediated transfer of AA.

  To characterize the transcriptional profiles of AA lesioned skin from C3H / HeJ mice and human AA, to identify genes with differential expression in skin between individuals with AA and skin from disease-free control individuals The present inventors performed Affymetrix microarray analysis. Three gene expression signatures have been identified in lesional skin: such as those encoding IFN-responsive genes such as IFN-induced chemokines CXCL-9, CXCL-10 and CXCL-11 in both human and mouse AA skin Transcription of several important CTL-specific transcripts, such as those encoding CD8A and granzymes A and B, and γc cytokines and their receptors, such as interleukin-2 (IL-2) and IL-15 Things. Since IL-2Rα has previously been shown to be expressed on infiltrating lymphocytes around human AA hair follicles, we have identified IL-15 and IL-15 and the source of IL-15 in the skin. Immunofluorescence analysis was performed for both of its chaperone receptor IL-15Ra. We detected significant up-regulation of both components in AA hair follicles in both human AA and mouse AA and found IL-15Rβ expressed on infiltrating CD8 + T cells in humans .

  IL-2 and IL-15 are well-known drivers of cytotoxic activity by IFN-γ producing CD8 + effector T cells and NK cells, and are involved in the induction and / or maintenance of autoreactive CD8 + T cells. In order to test the efficacy of IFN-γ- and γc-targeted therapy in vivo, we used a well-established graft model of AA. Here, skin grafts from mice with spontaneous AA are transplanted to the back of unaffected recipient C3H / HeJ mice at 10 weeks of age. In this model, AA develops reliably in 95-100% of transplanted recipients within 6-10 weeks, making it possible to test interventions aimed at preventing or reversing the disease.

  The role of IFN-γ in AA has been previously investigated using both knockout studies and administration of IFN-γ. In that case, the IFN-γ deficient mice were resistant and exogenous IFN-γ precipitation disease. Administration of neutralizing antibodies against IFN-γ at the time of transplantation prevented the development of AA in transplanted recipients and abrogated major histocompatibility complex (MHC) upregulation and CD8 + NKG2D + infiltration in the skin. (FIGS. 12A-12C). Similarly, the role of IL-2 in AA pathogenesis has previously been established using genetic experiments where IL-2 haploinsufficiency against the C3H / HeJ background confers disease resistance by about 50% using the graft model It had been. And this role is supported by genome-wide related studies in humans. Systemically administered blocking antibodies to either IL-2 (FIGS. 12D-12F) or IL-15Rβ (FIGS. 12G-12I) prevent AA in transplanted mice, prevent the accumulation of CD8 + NKG2D + T cells in the skin and up MHC Regulation was abrogated. However, IL-21 blockade failed to prevent the development of AA in transplanted C3H / HeJ mice. In particular, these blocking antibodies alone could not reverse established AA (data not shown).

  We next mediated the downstream effect of a small molecule inhibitor of JAK kinase that signals downstream of a wide range of cell surface receptors, allowing us to demonstrate the effect of blocking type I cytokines. We examined whether it could be reproduced. In particular, the IFN-γ receptor and the γc family receptor signal through JAK1 / 2 and JAK1 / 3, respectively. Activation of JAK is a phosphorylated signal transducer and activator of transcription (STAT) protein (pSTAT1, pSTAT3 and to a lesser extent pSTAT5) in human and mouse alopecia hair follicles. But was not shown in normal hair follicles. In in vitro cultured dermal sheath cells from C3H / HeJ mice, exogenous IFN-γ increased STAT1 activation, while IFN-γ + TNF-α increased surface IL-15 expression. Ruxolitinib, an FDA-approved small molecule inhibitor of JAK1 / J kinases (JAK selectivity is JAK1 = JAK2 = Tyk2 >> JAK3) important for IFN-γR signaling, is Inhibited these responses. In cultured CTL effectors from C3H / HeJ mice, the FDA-approved small molecule JAK3 inhibitor tofacitinib (JAK3> JAK1 >> JAK2 selectivity) blocked IL-15-induced pSTAT5 activation. Tofacitinib also blocked dermal sheath cell killing and granzyme B IL-15-induced upregulation and IFN-γ expression.

  In order to test whether inhibition of these signaling pathways is therapeutically effective in vivo, we transferred ruxolitinib (FIGS. 13A-13C) and tofacitinib (FIGS. 13F-13H) systemically at the time of transplantation. Administered. They found that they prevented the development of AA and the expansion of CD8 + NKG2D + T cells in all transplanted recipients.

  The skin of mice treated with either drug showed no signs of histological inflammation (FIGS. 13D and 131). A global transcriptional analysis of whole skin biopsy showed that both drugs were measured by an alopecia areata disease activity factor (ALADIN, FIGS. 13E and 13J) and gene expression dynamic index (GEDI) analysis. It has also been shown to block inflammatory signatures.

  The inventors then questioned whether systemic tofacitinib treatment could reverse the disease established by initiating treatment 7 weeks after transplantation. Systemic therapy resulted in substantial hair regeneration throughout the body, reduced the frequency of CD8 + NKG2D + T cells, and reversed histological markers, all of which persisted for 2-3 months after cessation of treatment .

  Next, to test a more clinically relevant delivery route, we established a local administration of protein tyrosine kinase inhibitors in mice with kinetics similar to systemic delivery kinetics. Asked if AA could be reversed. In established diseases, the inventors have found that topical ruxolitinib and topical tofacitinib are both very effective at reversing the disease in the treated lesion (applied to the skin on the back). It was. In mice treated with ruxolitinib or tofacitinib, full coverage of the hair appeared with 7 weeks of treatment. The inventors then observed complete hair regeneration within 12 weeks after local therapy (FIGS. 14A-14B). Topical therapy significantly reduced the percentage of CD8 + NKG2D + T cells in treated skin and lymph nodes (FIG. 14C), normalization of ALADIN transcription signature (FIG. 14D), and reversal of histological markers of disease (FIG. 14E). And associated with correction of GEDI in all treated mice. In particular, the untreated area of the abdomen remained alopecia (eg, FIG. 14A), demonstrating that the local therapy worked locally and that the observed therapeutic effect was not the result of systemic absorption. These effects were visible as early as 2 to 4 weeks after the start of treatment and persisted for 2 to 3 months after discontinuation of treatment (FIG. 14A).

  To test the effectiveness of JAK inhibitors in human subjects with AA, we treated three patients with mild to severe disease orally with 20 mg of luxolitinib twice daily. . Ruxolitinib is currently FDA approved for the treatment of myelofibrosis, a disease caused by wild type and mutant JAK2 signaling downstream of the hematopoietic growth factor receptor. Furthermore, small clinical studies with topical luxolitinib in psoriasis have demonstrated anti-inflammatory activity due to disruption of the IL-17 signaling axis. All three patients treated with ruxolitinib showed almost complete hair regrowth within 3-5 months of oral treatment (eg, FIG. 14F). A comparison of biopsies obtained at baseline and after 12 weeks of treatment showed reduced perifollicular T cell infiltration and reduced hair follicle expression of human leukocyte antigen class I and class II expression (FIG. 14G). ALADIN inflammation after placement and normalization of hair keratin signatures (FIGS. 14H and 14I) were demonstrated.

C. DISCLOSURE Taken together, the data suggests that CD8 + KG2D + T cells promote AA virulence while acting as cytolytic effectors responsible for hair follicle autoimmune attack. We have found that IFN-γ produced by CD8 T cells induces further production of IL-15 and a feedforward loop that promotes type I cell autoimmunity, while disrupting immune privileges in the hair follicle. Assume that it brings. The clinical response of a few AA patients to treatment with the JAK1 / 2 inhibitor luxolitinib suggests that future clinical evaluation of this compound, or other JAK protein tyrosine kinase inhibitors currently in clinical development, is warranted in AA ing.

D. Materials and Methods Mice The C3H / HeJ mouse strain (Jackson Laboratories, Bar Harbor, ME) was used for all animal studies. Only female mice were used. Mice recipients of alopecia skin grafts were 7-10 weeks old at the time of transplantation. For prevention experiments, drug administration was started the day after transplantation. For systemic treatment studies, drug administration began approximately 3 months after the mice lost their hair. For local treatment studies, drug administration began 20 weeks after transplantation. All animal treatments were performed according to protocols approved by the Columbia University Medical Center Institutional Animal Care and Use Committee.

Human studies All human studies were conducted under the principles of the Declaration of Helsinki, approved by the Institutional Review Board of Columbia University Medical Center. A consent form was received from participants prior to inclusion in the study.

Clinical evaluation of oral ruxolitinib in alopecia areata In our dermatology clinical trial unit at Columbia University Medical Center, we A single trial demonstration clinical trial entitled “Pilot Trial” was initiated (clinicaltrials.gov identifier: NCT01950780).

  The primary efficacy endpoint of this initial pilot study is the percentage of responders that achieved 50% or more regrowth at the end of treatment compared to baseline. Secondary endpoints include changes in hair growth both during and after treatment, measured as continuous variables; comprehensive assessment of patients; quality of life assessment; and persistence of response after treatment cessation.

  Inclusion criteria were 30-95% hair loss due to alopecia areata (AA) measured by SALT score; hair loss period of at least 3 months; stable hair loss without active evidence for regrowth; subject age 18-75 years Included.

  Exclusion criteria include active scalp diseases other than AA; medical history that may increase the risk associated with ruxolitinib, such as hematology, infectious, immune related disease or malignancy; may affect AA response Current treatment with any sex modality; drug treatment known to interact with ruxolitinib; pregnancy; etc.

  Study subjects will be treated with oral luxortinib 20 mg BID for at least 3 months. Patients in this document achieved over 90% regeneration. Skin punch biopsies (4 mm) were obtained at baseline and after 12 weeks of treatment.

Antibodies used for mouse treatment, flow cytometry, immunostaining and Western blot analysis All antibodies used in these studies are listed in the format of the following table.

  The following anti-mouse antibodies were used for flow cytometry analysis: CD3 (17A2, Ebioscience), CD4 (GK1.5, BD), CD8α (53-6.7, BD), CD8β (YTS156.7.7, Biolegend ), NKG2D (CX5, Ebioscience), NKG2A / C / E (clone 20d5, Ebioscience), CD44 (IM7, BD), CD45 (30-F11, BD), CD49b (Dx5, BD), CD62L (MEL-14) BD), CD69 (H1.2F3, BD), CD103 (2E7, eBioscience), IFNγ (XMG1.2, Ebioscience), Granzyme B (NGZB, eBioscience), Rae-1 (186107, R & D).

  For immunohistochemical studies of mouse skin, 8 μM methanol-fixed frozen skin sections were stained with primary rat antibody (Biolegend) including: anti-CD8 (clone 53-6.7), biotin anti-MHC Class I (clone 36-7.5), anti-MHC class II (clone M5 / 114.15.2). Biotin-labeled goat anti-rat IgG (Life Technologies) was used as the secondary antibody. For immunofluorescence studies, anti-H60 (R & D, clone 205326), anti-Pan Rae-1 (R & D, clone 186107), anti-KG2D (R & D clone 19004), anti-IL-15 (SCBT, H-114), anti-H IL-15 RA (SCBT, N-19), anti-K71 (Abeam), primary antibody was used in immunofluorescence. Alexa Fluor 488 or Alexa Fluor 594 conjugated goat anti-rat, donkey anti-rabbit or donkey anti-goat antibody was used as secondary antibody (Life Technologies).

  For immunohistochemical studies of human skin, 5 μM formalin fixed and paraffin skin sections were used. After heat antigen recovery, skin sections were stained with primary anti-human antibodies including: anti-CD8 (Abcam ab 4055), anti-CD4 (Leica clone 1-F6), HLA class 1 ABC (Abcam clone EMR8-5). ), HLA-DR / DP / DQ (SCBT clone CR3 / 43). ImmPRESS HRP anti-rabbit Ig or mouse Ig (peroxidase) polymer (Vector Lab) was used as secondary antibody.

  Human hair follicles were microdissected and embedded in OCT compound prior to sectioning and staining. 8 μM methanol-fixed frozen sections were prepared from IL-15 (SCBT, H-114) and anti-IL-15 RA (SCBT, N-19) or anti-IL-15 RB (SCBT, C-20) and CD8 (SCBT, C8). / 144B). Subsequently, staining was performed with Alexa Fluor 488 or Alexa Fluor 594-conjugated secondary antibody (Life Technologies). All images were taken with an SDRC Zeiss Exciter Confocal Microscope.

  For Western blotting, treated samples were resolved with 4-12% SDS PAGE (Life Technologies) and then transferred to Westran PVDF membrane (GE Healthcare Life Sciences). The blot was probed with the following antibodies (Ab) (all from Cell Signaling Technologies): anti-phosphoSTAT1 (Tyr701), anti-phospho-STAT5 (Tyr694), anti-STAT1 and anti-STAT5.

Antibodies for in vivo therapy

Antibody for flow cytometry (diluted 1/100)

Antibody for immunostaining and Western blot (1/100 dilution unless otherwise noted)

ＳＴＡＴ，ＳＴＡＴ５、ｐＳＴＡＴ１及びｐＳＴＳＴ５ab’はウェスタンブロットのために１／１０００希釈された。

STAT, STAT5, pSTAT1 and pSTATT5ab ′ were diluted 1/1000 for Western blot.

IL-15 and IL-15RA staining blocking reagent

RNA-Seq analysis Samples were sequenced over 50 cycles on a HiSeq 2000 sequencer (Illumina, San Diego, CA). The RNA-Seq file was demultiplexed by the Rockefeller University Genomic Score Facility. Quality control of the sample fastq file was performed using fastqe. TopHat was used to map the transcript from iGenome to the UCSC mm9 reference genome. RefSeq gene annotation packaged with this iGenome version of UCSC mm9 was used. The htseq-count utility from the HTSeq package was used to convert the ToppHat bam file into counts that could be used with edgeR as input for downstream analysis of differential expression. The missing gene was removed and a pseudo-count of 1 was added to avoid division by zero in the downstream analysis. EdgeR was used to identify genes with differential expression using a matched pair design with three biological replicates.

Microarray analysis Quality control, pretreatment For mice, cDNA samples were hybridized to a Mouse Genome 430 2.0 gene chip and subsequently washed, stained with streptavidin-phycoerythrin, and HP GeneArray Scanner (Hewlett-Packard Company). , Palo Alto, CA). In the case of humans, the amplified cDNA was hybridized to a human genome U133 plus 2.0 gene chip.

Microarray quality control and pretreatment were performed at R using a BioConductor. 3 experiments,
1) Spontaneous AA mice versus normal mice,
2) Prophylactic mice vs. placebo and sham-operated mice with 3 treatments, and 3) Treatment mice vs. placebo for 2 treatments,
Were pre-treated separately using the same pipeline.

  Quality control is available at http://arrayanalysis.com. org / affyanalysis QC package was used. Affyanalysis QC uses R / BioConductor package, ie, affy, affycomp, affydnn, affyPLM, affyQCReport, ArrayTools, bioDistm bioty, affy, affy, affy, ff, ff RMA normalization was performed separately for each experimental group. The prevention experiment required correction of batch effects using ComBat. Batches, treatments and time points were modeled by treating each treatment group effect as constant over time and grouping PBS controls in groups that reflected both treatment and time.

  In addition to the pretreatment performed on mouse skin samples, Harshlight was used to correct image defects in human skin samples.

Deposition of data
Microarray and RNA-seq data were deposited at Gene Expression Omnibus (GEO), accession numbers GSE45657, GSE45512, GSE45513, GSE45514, GSE45551 and GSE58573.

Identification of gene signatures Differential expression analysis
Initial analysis of differential gene expression was performed on spontaneous mouse 3 × 3 and human 5 × 5 data sets using limma. 1.5 fold change threshold and 0.05 unadjusted p-value.

Unsupervised analysis Hierarchical clustering is performed on human 5 × 5 derived 363 gene and spontaneous mouse 3 × 3 derived 583 gene satisfying threshold abs (logFC)> 1, unadjusted p-value <= 0.05 This was done using clusters. Genes were median centered and normalized. Spearman rank correlation was used as a similarity measure, and an average ligation method was used for clustering of rows (genes) and columns (samples). Visualization of hierarchical clusters was performed using Java® TreeView. Gene expression dynamic index (GEDI) analysis was used to visualize how the “metagenes” identified by the self-organizing map algorithm differ from sample to sample. A metagene is a cluster of genes that shows a similar expression pattern across a sample and is assigned to a single pixel in a two-dimensional grid.

RT-PCR validation RT-PCR was used to confirm genes with predicted differential expression in humans and mice. First-strand cDNA was synthesized using a ratio of 2: 1 random primers (Oligo (dT) primer and SuperScript III RT (Invitrogen)) according to the manufacturer's instructions. qRT-PCR was performed on an ABI 7300 instrument and analyzed with ABI relative quantitative test software (Applied Biosystems, Foster City, CA, USA). Primers were designed according to ABI guidelines and all reactions were performed using Power SYBR Green PCR Master Mix (Applied Biosystems), 250 nM primer (Invitrogen) and 20 ng cDNA in a 20 μL reaction volume. The following PCR protocol was used: Step 1: 50 ° C. for 2 minutes; Step 2: 95 ° C. for 10 minutes; Step 3: 95 ° C. for 15 seconds; Step 4: 60 ° C. for 1 minute; Repeat cycle. All samples were repeated 4 times for 3 independent runs and normalized to the endogenous internal control as indicated.

ALADIN score IFN and CTL signatures were used to develop bivariate score statistics. Individual signature IFN and CTL scores were determined according to the procedure used for human SLE. The set of genes selected to include IFN and CTL signatures were CD8A, GZMB and ICOS for CTL signatures and CXCL9, CXCL10, CXCL11, STAT1 and MX1 for IFN signatures. Prophylactic mouse scores were calculated for sham mice, while local treatment experiment scores were calculated for all samples at week 0. Based on human studies, ALADIN has been further extended to include hair keratin (KER) signatures. The set of genes selected to contain the KER signature is DSG4, HOXC31, KRT31, KRT32, KT33B, KRT82, PKP1 and PKP2. Baseline and 12-week skin biopsy ALADIN scores obtained from subjects enrolled in the oral ruxolitinib clinical trial were calculated relative to healthy controls at baseline.

Power analysis For the analysis of response to treatment, we have determined that the true proportion of population 1 (treatment group) expected to respond to treatment is 0.95 and the response If the true proportion of population 2 (placebo group) expected is 0.20, compare the two samples for the ratio group power size for the treatment group and placebo mice, for each of the five group sample sizes Went. At the significance level α = 0.05, using Barnard's exact test, we have two population proportions of 0.95 and 0.20, and the group sample size is equal to 5 each. In addition, a power of 0.803 was calculated for a one-sided test to detect the difference in ratio. In some instances where there were less than 5 animals per group per experiment, multiple experiments were disrupted to ensure statistical power.

Statistical analysis of therapeutic effect Mice were expected to develop alopecia after 4-12 weeks of transplantation of hair loss skin. Experiments in which control mice did not show hair loss until 8 weeks were discontinued. For prevention experiments, a time-to-event survival analysis was performed on interval censored data. R survival and interval packages were used to perform the log rank test. The hair growth index was calculated.

  For the treatment experiment (FIG. 14B), the R package nparLD was used to test the hypothesis that treatment with time interaction exists. The analysis was performed using the hair growth index from three replicates including a placebo group of 3 mice from each treatment and a total of 9 mice from each group. F1-LD-F1 design was adopted. In the case of JAK1 / 2i treatment vs. placebo, the no interaction hypothesis, ie parallel time profile, is rejected at the 5% level using both Wald-Type and ANOVA-Type statistics. Here, the p values are 4.40e-21 and 3.35e-18, respectively. In the case of JAK3i treatment vs. placebo, the no interaction hypothesis, ie parallel time profile, is rejected at the 5% level using both Wald-Type and ANOVA-Type statistics. Here, the p values are 1.45e-30 and 2.42e-21, respectively.

  All mice were included in survival (time vs. event) analysis statistics. For lymph node and skin cell analysis, biopsies were taken at the indicated time points after treatment in parallel with control mice. In the IFN-γ and IL-2 neutralization experiments, one of the 5 control mice that did not show hair loss was not included in the photograph. Although these mice were not sacrificed to continue monitoring hair loss, for statistical purposes of skin cell analysis, these unanalyzed samples included strict and conservative statistics with treated mice. A cell count value of 0% CD8 + NKG2D + cells was assigned to allow for comparative comparison.

  Randomization was not used and researchers were not blinded to group assignments during the experiment or when evaluating results.

  Mean and frequency differences between the treated and untreated groups were tested using an unpaired parametric two-tailed t-test. For statistical purposes, we assume that all variances are the same for each group.

  To perform all time to event survival analysis, a period-censored log-rank test was used. This test adequately accounts for data where the exact event time is not known but the event is known to be within an interval.

  Non-parametric longitudinal data analysis was used to test for response x time interaction. These methods are particularly suitable for small sample sizes.

Sample size, number of replicates, and statistical tests between experiments

  For in vivo studies, data is provided as cumulative data. The number of replicas is provided as described above. For example, “3 + 3 + 3 + 3” refers to 4 separate experiments involving 3 experimental mice. Experiments were performed in triplicate for in vitro studies.

6.3 Example 3
Summary Alopecia areata (AA) is a very frequent autoimmune disease in the United States with a lifetime risk of 1.7%. However, AA continues to be an unmet need in dermatology and a cure is lacking. Janus kinase inhibitors have emerged as potential therapies for many autoimmune conditions, most recently including AA. We have taken oral tofacitinib citric acid, a preferential JAK3 / JAK1 inhibitor, for AA resulting in significant hair regrowth and concomitant changes in skin and blood biomarkers. Report patients treated with salt. We hypothesized that effective tofacitinib treatment of alopecia areata would be accompanied by altered expression of AA-related genes in the skin as well as circulating serum CXCL10 levels. Punch biopsies were taken at baseline and after 4 weeks of treatment. Total RNA was extracted, reverse transcribed, amplified, biotinylated, and then hybridized to a Human U133 Plus 2.0 gene chip (Affymetrix, Santa Clara, Calif.). The ALADIN score was calculated as described above compared to healthy controls at baseline. Sera from blood draws collected before the start of tofacitinib treatment and after 4 weeks of treatment were collected using the Human IP-10 / CXCL10 ELISA kit (Sigma-Aldrich, St. Louis, MO) according to the manufacturer's instructions. Used and evaluated for CXCL10 levels according to manufacturer's instructions.

Results A 40-year-old white female with persistent moderate / severe AA was enrolled in an open-label pilot study to test the efficacy of oral tofacitinib against AA (https://clinicaltrials.gov/NCT02299297 ). Her AA began on her scalp five years before enrollment and was completely resolved in pregnancy settings within one year. A few months after delivery, her AA recurred as a patchy disease. Treatment with topical corticosteroids, anthralin cream and intralesional corticosteroids had limited benefits. Her AA progressed to the spread of mottled scalp across all limbs, eyelashes and eyebrows and remained stable until her enrollment in clinical trials (FIG. 15). Her past medical history is mediocre, and she denied AA's family history.

  The patient began treatment with 5 mg of tofacitinib twice daily. Spot-like regeneration was observed in the first month. After 2 and 3 months of treatment, her scalp hair regeneration was 62.5% and 94%, respectively. Significant regeneration of her eyebrows and eyelashes was observed. The scalp hair regeneration was almost complete 4 months after the start of treatment (FIG. 15). No adverse events were reported and there were no laboratory abnormalities in her whole blood count, complete metabolic panel or lipid profile. Discontinuation of treatment with tofacitinib resulted in almost complete hair loss (FIG. 16).

  Scalp punch biopsy (FIG. 17) and blood collection were performed at baseline and after 4 weeks of treatment to perform blood collection and to monitor gene expression and biomarker changes. Serum levels of CXCL10, an interferon (IFN) -induced chemokine found at high levels in AA skin, decreased after 4 weeks of treatment (FIG. 17). In addition, microarray analysis was performed on skin biopsy samples. Based on the AA disease activity index (ALADIN), patients show high IFN and cytotoxic T lymphocyte (CTL) signatures at baseline, which is not reduced to normal control levels but decreased by 4 weeks of treatment (FIG. 17).

Conclusion Alopecia areata is an autoimmune disease with a strong association with loci that are in close proximity to genes with immune function. Targeting candidate immune pathways that can contribute to disease pathogenesis is an active area of research, and JAK inhibitors target multiple immune signaling pathways involving AA. We have shown that in a transplantation model of AA in C3H / HeJ mice, systemic and local tofacitinib is effective in preventing the development of AA and reversing established AA. Shown earlier. We now report an effective treatment of a human subject with persistent mottled AA that correlates with a reduced ALADIN profile compared to baseline.

6.4 Example 4
Summary Alopecia areata (AA) is a common autoimmune disease with a lifetime risk of 1.7% and there is no FDA approved treatment for alopecia areata (AA). We have previously identified a dominant IFNg transcription signature within cytotoxic T cells (CTL) in human and mouse AA skin, and treatment with JAK inhibitors in mice by targeting this pathway. It has been shown to induce durable hair regeneration. Here, the inventors studied the use of the oral JAK1 / 2 inhibitor luxolitinib in the treatment of patients with moderate to severe AA.

  We started an open-label clinical trial of 12 patients with moderate to severe AA using oral ruxolitinib 20 mg BID over 3-6 months of treatment, followed by 3 months of follow-up drug withdrawal (Follow-up off drug) followed. The primary end point was the proportion of subjects with 50% or more hair regeneration from baseline to the end of treatment.

  Nine of the 12 patients (75%) showed a significant response to treatment, with an average hair regeneration of 92% at the end of treatment. Most of the safety parameters remained within normal limits and no serious adverse effects were reported. Gene expression profiling revealed treatment-related down-regulation of inflammatory markers, including signatures for CTL and IFN-responsive genes, as well as upregulation of hair-specific markers.

  In this pilot study, 9 out of 12 patients (75%) treated with ruxolitinib showed significant scalp hair regeneration and improved AA. Larger randomized controlled trials are needed to further evaluate the safety and efficacy of ruxolitinib in the treatment of AA.

Introduction Alopecia areata (AA) is an important medical problem, the most frequent autoimmune disease in the United States, with a lifetime risk of 1.7%. AA affects both sexes across all ethnic groups and represents the second most common form of human alopecia following androgenic alopecia alone. AA usually shows patchy hair loss. One third of these patients will experience spontaneous remission within the first year. However, many patients' disease will progress to total alopecia (AT, total scalp alopecia) or generalized alopecia (AU, loss of whole body hair). Persistent, moderate / severe AA causes significant appearance effects and mental distress for affected individuals. In clinical practice, there is no evidence-based treatment for AA, but various treatments are provided, most commonly topical steroids and intralesional steroids, which have limited efficacy It has.

  Recent studies have demonstrated a dominant role of type I cellular immunity in the pathogenesis of AA, mediated by CD8 + cytotoxic T lymphocytes (CTL) with NKG2D producing interferon gamma. The central role of type I cellular immunity is also reflected in the transcriptional landscape of AA lesion skin in humans and mice, governed by IFN response genes and CTL signatures. These findings provide a rationale for the therapeutic targeting of JAK1 / 2 kinase in AA, in fact, we found that treatment with JAK inhibitors reversed AA in C3H / HeJ mice. It has been shown to eliminate type I inflammatory responses in the skin.

  Based on preclinical findings, the inventors have initiated a phase 2 efficacy signal pursuit clinical trial in moderate to severe AA and currently have JAK1 / 2 approved by the FDA for the treatment of myeloproliferative disorders The clinical and immunopathological response to treatment with the inhibitor, oral ruxolitinib, was evaluated.

Methods Study design, supervision, and participants The study was devised and carried out by a research team at Columbia University. All authors had access to the data, demonstrating its accuracy and fidelity to the research protocol of this report. This study follows the Good Clinical Trial Standards (GCP), as defined by the International Conference on Harmonization (ICH), and is part of the European Union Directive 2001/20 / EC and US Federal Regulations, Title 21, Part 50 (21 CFR50) Implemented in accordance with the underlying ethical principles. Prior to the start of the study, the protocol and all research-related materials were approved by Columbia University IRB. Prior to screening or study-related procedures, free written informed consent was obtained from all subjects. Monitoring for regulatory compliance and adherence to IRB approved protocols was performed by the Department of Columbia University Clinical Laboratory and Surgical Rules Team. The study was conducted before clinicaltrials. registered in gov. We enrolled 12 adult patients, including 10 patients with moderate to severe AA (30-95% hair loss) and 2 patients with AT or AU.

Research evaluation and prognosis (Outcomes)
The primary efficacy endpoint of the study is at least 50 compared to baseline as assessed by alopecia severity tool (SALT) score, a method for estimating hair loss in standardized and validated AA. The percentage of responders at the end of the treatment, defined as those subjects achieving% regeneration. Secondary efficacy endpoints included hair regeneration as a continuous variable. In addition, lifestyle measurements (dermatological quality of life indicators-DLQI and Skindex) were performed at regular pre-specified intervals, but showed no statistical differences in the comparisons performed (data not shown) It has not been). To assess response durability, the responder was followed for 3 months after the treatment was completed. Safety analysis was included as a secondary endpoint for all subjects who received at least one dose of ruxolitinib and were observed at the monthly visit as described.

Exclusion criteria Patients with AA less than 3 months; Active, unstable AA or regenerating AA; Combinations that may affect hair regeneration (within one month prior to enrollment) A patient was excluded if he had received treatment; or if he had evidence of an underlying infection, malignancy, immunodeficiency or unstable medical condition. Also excluded were patients with skin disease associated with the scalp; or patients receiving experimental drug therapy within the last month or three times the drug half-life. Patients who reported recently or use of DMARDs (disease modifying anti-rheumatic drugs) were excluded.

Adverse effects Adverse events may be any new untoward medical consequences in patients who received at least one dose of study drug, regardless of whether the event was considered causal to the study treatment. Classified as an event (signs, symptoms or abnormal laboratory findings), or a deterioration of an existing medical condition. Adverse events were assessed at the monthly visit. Patients were also encouraged to contact the research center during the visit if they developed new signs or symptoms of concern. Several patients initially developed a gradual decrease in white blood cell counts, but the levels remained within normal limits and therefore no dose adjustment was required. One patient developed reduced hemoglobin levels, which required a dose change. No significant reduction in platelet count was observed. One patient developed two episodes of rash / abscess. Both episodes were evaluated by the patient's attending physician and resolved before the patient was evaluated by the research team. The patient also reported possible biopsy site infections, for which she was reported seeking medical attention and being treated with oral antibiotics abroad. Some patients developed mild URIs that appear seasonal and unrelated to the drug. One patient had a mild episode of pneumonia confirmed by chest x-ray, which was treated with oral antibiotics. This patient had a distant history of pneumonia episodes several years prior to study participation. No clinically significant reduced platelet development was observed. There were no liver abnormalities. One patient had elevated lipids at baseline and during treatment. He was monitored by his attending physician while on study medication and had no clinically apparent adverse effects related to lipid levels. Two patients developed lesions consistent with acne or scalp folliculitis. Both episodes were resolved in a few weeks. Three patients had GI symptoms including diarrhea. One patient developed conjunctival bleeding (a common side effect of the procedure) and transient hemorrhoid hemorrhage following a pre-planned ophthalmic procedure. There was no change associated with circulating levels of platelets. Some patients had mild abnormalities in urinalysis / microscopy. One patient received treatment for a urinary tract infection.

Biomarker assessment and clinical correlation studies Biopsies and peripheral blood were obtained at baseline and after 12 weeks for immune surveillance and molecular studies. Some patients provided an additional biopsy at an intermediate time during the course of treatment, and one patient provided an additional sample at 24 weeks. Tissue specimens were fixed and stored in PAXgene Tissue Containers. Total RNA was extracted from skin biopsy specimens taken during the course of clinical trials using the PAXgene tissue miRNA kit. Library preparation was performed for microarray analysis using Ovation RNA Amplification System V2 and Biotin Encore Kits (NuGen Technologies, Inc., San Carlos, Calif.). Subsequently, the samples were hybridized to a human genome U133 plus 2.0 chip (Affymetrix, Santa Clara, Calif.) And scanned at Yale Center for Genome Analysis. Library preparation of RNA extracted from skin biopsies from three healthy controls and microarray hybridization were performed on a total of 31 samples along with samples from treated patients. Gene expression analysis included calculation of the ALADIN score, expression level differential expression analysis for gene expression signature identification, principal component analysis, and statistical analysis of the ALADIN score. Microarray data from 31 samples has been deposited with GEO under the accession number GSE80342.

Microarray Pretreatment and Quality Control Microarray quality control and pretreatment was performed using a BioConductor at R. Quality control is available at http://arrayanalysis.com. This was done using the R stand-alone version of affyAnalysis QC from org.

  Samples included in 31 from this study, plus three additional samples from healthy controls from GEO accession number GSE53573, were normalized together using GCRMA. The Affymetrix probe set was called present or absent by affyAnalysis QC using the MAS5 algorithm implemented in R. Probe sets IDs (PSIDs) that were on the X or Y chromosome, PSIDs that were Affymetrix control probe sets, or PSIDs without the Gene Symbol annotation were removed from all arrays for further downstream analysis. Data was analyzed using log2 (Intensity) for expression levels.

  A batch effect correction was required to include three healthy control samples from the previous data set GSE58573 for calculation of the ALADIN score and for improved increased power in differential expression analysis. A modified version of ComBat (M-ComBat) was used to integrate normal samples from previous GEO datasets. While M-ComBat was used to integrate the three healthy control samples from GSE58573 with the current healthy control sample, the expression level of the sample from the current data set was held fixed. M-ComBat is an implementation of the functionality that ComBat is available in the sva package that allows one of the batches to be used as a reference batch.

Calculation of ALADIN score The ALADIN score was calculated for all 34 samples using batch-corrected expression data. CTL, IFN and KRT ALADIN scores were determined according to the procedure outlined above. The z-score is briefly calculated for each PSID compared to the mean and standard deviation of normal controls. The score for each gene is obtained by averaging the z-scores of the PSIDs mapping to that gene. The signature score is then the calculated average of the scores for the genes belonging to the corresponding signature.

Identification of gene expression signatures To perform further analysis on baseline luxolitinib samples (n = 12) along with normal controls (n = 6), PSIDs were called to be present in at least one of 18 samples Further filtering to retain only the features, 36147 PSIDs remained for further downstream analysis.

Filtering Paired Samples at Times t = 0 and t = 12 from Patients Responding to Ruxolitinib Treatment After QC, 8 with microarray data at both t = 0 and t = 12, in response to ruxolitinib treatment There were patients. The PSIDs were further filtered to retain only features that were called to be present in at least one of the 16 samples, leaving 35563 PSIDs for further differential expression analysis.

Differential expression analysis Differential expression analysis was performed on samples from t = 0 and normal controls, and from responders, from t = 0 and t = 12, using a linear model implemented in the Bioconductor limma package. Executed on paired data. A 2.0 fold change threshold and an unadjusted p value of 0.05 were used.

  Comparisons between responders at baseline and normal controls, between responders and non-responders at baseline, between non-responders at baseline and normal controls, responders at baseline and responders at 12 weeks of treatment As well as between responders and normal controls in the last 12 weeks of treatment.

Principal component analysis of 6 healthy control and luxolitinib treated patients at t = 0 and t = 12. Principal component analysis is a responder signature consisting of PSIDs differentially expressed between responders and normal controls at baseline. Was performed on samples from 6 healthy controls and ruxolitinib treated patients at baseline and 12 weeks.

Statistical analysis All variables were examined for distribution assumptions and checked for accuracy and out-of-range values. Based on a priori definition, we have experienced 50% or more hair regeneration from baseline based on the SALT score at the end of treatment. The patients were classified as responders. We looked at the overall distribution of demographic factors, examined possible differences between responders and non-responders, and Fisher's Exact Test for categorical variables ( Bilateral) was used to test for significance using the Mann-Whitney U test for continuous variables. The inventors then employed either the Mann-Whitney U test or the Wilcoxon Signed Rank test to determine the baseline for all relevant variables and for responders and non-responders. The change between end of treatment (EOT) and end of study (EOS) scores was examined. In order to estimate the degree of regeneration over time, we consider both the Generalized Estimating Equation and the mixed model approach for modeling repeated measures data, and the strong normality of GEE The latter was chosen on the assumption of (normality) and a relatively small number of cases. For these mixed models, we first modeled regeneration from baseline to the end of treatment, where time (weeks) was the independent variable. The inventors then modeled regeneration from the end of treatment to the end of the study to evaluate the maintenance of the observed effect, where time was again an independent variable. In both models, we identified compound symmetry as the initial covariance structure.

Statistical analysis of ALADIN score In order to examine possible differences between CTL, IFN and KRT ALADIN scores in responders and non-responders, responders and normal controls and non-responders and normal controls, we We tested for differences between the three groups using the Kruskal-Wallis test followed by the Wilcoxon Rank Sum tests as implemented in the coin package in R. We then examined the change between the baseline and the ALADIN value at 12 weeks for the responder using the Wilcoxon sign rank test as implemented in the coin package in R. We further tested for a three-score difference between responders and normal controls at 12 weeks.

The ALADIN score is defined such that the average CTL, IFN, and KRT scores are equal to zero, resulting in an average overall (all patients) score, responder-only score, and non-responder-only score, and the difference between these average and normal Corresponds to the control. Statistically significant differences were observed in the CTL (p <0.0002), IFN (p <0.005) and KRT (p <0.0002) scores between the overall score at baseline versus normal controls, Responder only at baseline vs. normal control was observed in CTL (p <0.0004), IFN (p <0.0004), and KRT (p <0.004); overall score vs. normal control at 12 weeks Between CTL (p <0.04) and IFN (p <0.0004);
Between responder only vs normal control, alpha = 0.05 was observed in KRT (p <0.0007). No statistically significant differences were observed in IFN scores in baseline versus normal controls at baseline; or in responder-only vs. CTL and IFN scores in normal controls at 12 weeks. No statistically significant difference was observed between nonresponders at baseline and normal controls in any of the three ALADIN scores.

Between baseline and 12 weeks, changes in ALADIN scores in individual patients were evaluated. Statistically significant differences are observed in CTL and IFN scores between the overall baseline and week 12, with KRT scores reaching marginal significance (α = 0.05) there were. The CTL score decreased from 8.30 to 1.51 (p <0.004), the IFN score decreased from 31.08 to −0.37 (p <0.004), and the KRT score was −39.36. To -15.02 (p = 0.054). Among responders only, the CTL score decreased from 9.37 to 1.6 (p <0.008), the IFN score decreased from 38.37 to 0.24 (p <0.008), and the KRT score Increased from -37.84 to -15.42 (p = 0.039). Statistically significant differences were observed between responders and non-responders in CTL and IFN scores at baseline (mean score differences = -5.91 and 40.11, p <0.036 and 0, respectively) 0.036) was not observed in the KRT score (mean score difference = −19.74, p = 0.22).

Results Efficacy This study was an open-label clinical trial to investigate 20 mg PO, twice daily luxolitinib (Jakafi, Incyte Pharmaceuticals) in the treatment of moderate / severe AA. All patients received ruxolitinib for 3-6 months followed by a 3-month observation phase to assess treatment response endurance.

  Nine of the 12 patients (75%) had significant hair regeneration and achieved a major prognosis of at least 50% regeneration. The average baseline SALT score of 65.8 ± 28.0% dropped to a score of 24.8 ± 22.9% at 3 months and 7.3 ± 13.5% at the end of 6 months of treatment (P <0.004). As one group, responders showed a 92% reduction in hair loss from baseline (FIGS. 18 and 19) and 7 out of 9 responders achieved more than 95% regeneration by the end of treatment. .

  Regeneration is seen in the responder 4 weeks after the start of study medication, showing mainly gray hair regeneration, with the exception of one patient (subject 4) who developed vitiligo, pigmented terminal hair It initially appeared as a subtle patchy region of variability in regeneration. Of note, the area of vitiligo in this patient was also observed to improve with ruxolitinib treatment. All responder hair regrowths steadily increased with a significant increase every month, resulting in the majority of responders (8 out of 9) achieving at least 50% regrowth by the 12-week visit. It was. Responding patients with evidence of regeneration at 3 months continued treatment until the subject achieved 95-100% regeneration or completed 6 months of treatment.

  Response endurance was assessed within a 3-month follow-up period from treatment. Three of the nine responders observed the onset of shedding at 3 weeks following discontinuation of ruxolitinib and had significant hair loss at 12 weeks after dosing (Figure 21); however, hair loss was at baseline levels Did not reach (FIG. 18). Six of the nine responders reported a slight increase in dropout.

Biomarkers and clinical correlation studies Gene expression profiling is performed at baseline on skin biopsies taken following 12 weeks of treatment, and additional, selective biopsies are performed early in the course of treatment. It was done. Baseline scalp samples showed distinct gene expression profiles when compared to samples taken from unaffected patients (FIG. 20A). Following ruxolitinib treatment, AA patient scalp samples gathered more closely in healthy control scalp samples than baseline AA samples, indicating an overall normalization of the AA pathogenic response (FIG. 20B). Gene expression profiles attributed to interferon (IFN), cytotoxic T lymphocyte (CTL), and hair keratin (KRT) signatures are trivariate alopecia areata, a summary indicator of AA pathogenic inflammatory responses and hair regeneration The disease activity index (ALADIN, FIG. 20C) was evaluated. Importantly, the final AA responders gathered together on the ALADIN matrix at baseline and shared high IFN and CTL scores (FIGS. 20C, 20D).

  Of note, baseline samples from the final AA non-responder showed relatively low IFN and CTL scores that were not statistically different from normal control samples (FIGS. 20D, 20E). In addition, in a cohort of AA patients for ruxolitinib treatment as described, CTL and IFN signature scores were able to distinguish between final responders and non-responders at baseline (for CTL and IFN scores, respectively) p <0.036 and p <0.036).

  Consistent with on-target activity, skin samples taken from responding patients following 12 weeks of treatment show much lower IFN and CTL scores, and normal controls on the ALADIN matrix Much closer gathered in skin samples taken from patients (FIGS. 20D, 20E). Decreased IFN and CTL scores in post-treatment biopsies were evident as early as 2 weeks after the start of treatment (FIG. 22).

Adverse events Luxolitinib was well tolerated and administered safely in all 12 patients. There were no serious adverse effects and no patients were required to discontinue treatment. Observed adverse effects are rare, 3 mild bacterial skin infections (in the same patient), 9 episodes of URI / allergic symptoms in 7 patients, 1 UTI, 1 mild pneumonia A mild GI symptom, and a conjunctival hemorrhage following a surgical procedure. One patient developed hemoglobin decline and was resolved by a dose change.

Discussion In this empirical study, 20 mg of ruxolitinib twice daily for 3 to 6 months induced significant hair regeneration in 9 of 12 patients, and in the treatment of AA, overall against luxritinib The response rate was 75%. In contrast, the expected spontaneous remission rate (occurrence of hair regeneration without treatment) in patients with moderate to severe AA is shown in two randomized controlled trials in a similar population of subjects. ) Is less than 12%. Even the most severe form of alopecia, AT / AU, responded, indicating that the autoimmune process remains pathogenically active and is reversible by JAK inhibition. Hair regeneration was evident within one month in the responder and proceeded rapidly. The response was almost complete by 6 months of treatment in 8 out of 9 responders, suggesting that 6 months of treatment is sufficient to induce maximal clinical remission in the majority of responders.

  In this 9-month study, ruxolitinib was well tolerated. The safety signal in this small study of otherwise healthy AA patients, especially in patients with myeloproliferative diseases where it is understood that hematologically related adverse events are more frequent, It is superior to clinical experience and is consistent with the findings from the use of tofacitinib in the treatment of patients with psoriasis.

  Transcript profiling of paired baselines and scalp biopsies during treatment was both mechanistic and clinically informative. Baseline skin samples from responders had highly inflammatory ALADIN IFN and CTL scores, but almost normalized after 12 weeks of treatment, indicating JAK1 / 2i-mediated suppression of autoreactive CD8 T cell responses. In fact, early ALADIN normalization (FIG. 22) as early as 2 weeks following initiation of treatment may predict a favorable 12th week clinical prognosis. Conversely, non-responder samples show a low baseline IFN / CTL score and are relatively closely clustered with normal patient samples, an alternative inflammatory or non-inflammatory of hair loss in these non-responders Suggests etiology (Figure 23). One non-responder has both AA and androgenic alopecia, and the other non-responder alopecia was histologically consistent with AA, but appears to be a rare diffuse form of the disease. The non-responder showed a serpentine hair loss AA pattern.

  A recent single case report describes the clinical response in AA patients treated with other JAK inhibitors including tofacitinib, ruxolitinib, and varicitinib. These empirical data demonstrate the immunopathological reversibility of the type I inflammatory response underlying AA, even in patients with long-term or more severe disease forms, and oral and / or for treatment of AA Or provide a strong rationale for the clinical development of topical JAK inhibitors.

6.5 Example 5
Summary Although research on complex diseases such as cancer has proved very important, network-based molecular modeling of physiological behaviors, these approaches involve situations involving interacting tissues such as autoimmunity. Most of the (contexts) remain untested. Here, alopecia areata (AA) is used as a model, and we have performed a control network analysis to specifically isolate physiological behaviors in the skin that contribute to immune cell mobilization in autoimmune diseases. Adapted. We use context-specific control networks to deconvolve and identify skin-specific control modules with IKZF1 and DLX4 as the main regulators (MRs). Is used. These MRs are sufficient to induce an AA-like molecular state in vitro (in vitro) in three cultured cell lines, resulting in induced NKG2D-dependent cytotoxicity. This work demonstrates the feasibility of a network-based approach to compartmentalize and target molecular behavior that contributes to tissue interactions in autoimmune diseases.

Introduction System-level analysis using retrospectively analyzed control networks is a new computational discipline that has shown great promise in the study of complex diseases such as cancer and Alzheimer's disease. This approach allows the modeling of complex physiological behavior as a module of genes (a subset of differentially expressed genes associated with disease) controlled by key regulators (MRs). MRs represent the physiological behavior associated with the minimal number and expansion of transcription factors (TFs) that are predicted to specifically activate or repress target modules. They can be viewed as molecular “switches” that regulate physiological behavior. Inference of MRs is made possible through retrograde analysis of situation specific control networks using computational algorithms such as ARACNe.

  These MRs are biologically validated and serve as targetable “hubs” that govern the pathology of the disease. These approaches have proven to be highly effective for studying cell autonomous behavior in diseases such as cancer. Physiological behaviors such as mesenchymal transformation in glioblastoma and tumor development in B cell lymphoma or breast cancer, and the onset of Alzheimer's disease are functionally linked to a relatively small number of MRs, which in turn is It becomes a “bottleneck” that can be used to infer driver mutations, or a target for drug screening for treatment.

  However, this type of computational approach is only beginning to be implemented to target pathogenic, non-cellular autonomous interactions between different tissues, such as autoimmune diseases. In particular, inferring MRs cannot be done directly using typical ARACNe-based analysis due to underlying assumptions made during control network generation: (1) Samples used Represents a relatively pure or one underlying transcriptional network; and (2) the molecular behavior underlying the data set exists in a steady state, so that each sample has control dependence within the entire network. May be treated as a “snapshot”. Samples that are contaminated by tissues that exhibit particularly different situation-specific control networks can compromise the accuracy of control predictions. Furthermore, if pathogenesis depends on the interaction between two tissues, there will always be an artifact correlation between contaminating gene signatures and molecular modules that mobilize them but are expressed in other tissues. This makes it difficult to clearly define a module that is exclusive to one tissue or the other when analyzing gene expression data generated from a mixture of two tissues.

  Alopecia areata is ideal for such studies because it is characterized by cytotoxic follicles that actively infiltrate hair follicles and scalp skin that are typically absent from normal skin. The right model. AA typically manifests as the loss of individual random mottled hair, which can spread throughout the scalp (total alopecia) or systemic (generic alopecia). Previous studies have directly linked immune genes to AA, many of which are shared with other autoimmune diseases such as type 1 diabetes, celiac disease and rheumatoid arthritis. Previous studies have identified the invasion of cytotoxic CD8 positive, NKG2D positive T cells into the skin of AA, and the pathology of AA is associated with an IFN-gamma-dependent signaling pathway, which in cancer Often disrupted with immune evasion.

  Determine whether intrinsic factors are present in “end organs” (tissues subject to autoimmune attack) that contribute to the disease, such as scalp skin in AA, and analyze this molecular component as the primary target for analysis There is little work to do. We expect that pathogenic changes in the molecular profile of the scalp skin include genes that mediate interaction with infiltrating T cells. Inevitably, identifying MRs will provide sufficient regulatory control to the module to induce immune recruitment. To study this, we utilize a situation-specific control network for regulatory deconvolution of the mixed signature gene expression profile of AA patients. The goal of this work was to develop a framework that could separate the mixed AA tissue biopsy gene expression data into skin-specific modules of AA pathology and infiltration mobilization.

  We have identified a molecular profile of AA containing a gene module for invasive mobilization in scalp skin by filtering genes that do not map correctly to the skin-specific network. This scalp skin signature allowed the subsequent identification of the two MRs of scalp skin: IKZF1 and DLX4 to contribute to invasion. These two genes are expressed in major scalp biopsies and are sufficient to induce AA-like molecular signatures and NKG2D-dependent cytotoxicity in an independent wild-type cellular context, direct genetics of immune-mediated cytotoxicity Enable guidance.

Results The initial definition of pathogenic expression signatures in AA reveals the presence of local scalp skin and infiltrating immune signals. First, we compared the AA patients to the molecular expression of AA compared to controls ( A molecular signature that generates a molecular representation was created. We analyzed a training set of microarray studies of patient biopsies from the first cohort of 34 unique biopsy samples: 21 AA patients in various clinical presentations, and 13 affected No control. In addition, we performed patient-matched, non-lesional scalp biopsies for 12 of 21 AA patients. These 34 patients were collected as the first of two cohorts of 96 patients in total, and the rest were saved for validation studies.

  We have compared the global gene by comparing two separate clinical presentations, mottled AA (AAP) and complete and generalized (AT / AU) patients to unaffected controls. An expression signature was created. To account for signature artifacts associated with secondary effects of invasion, such as hair loss, we next looked at patient-compatible lesions (symptomatic skin with hair loss) samples and non-lesions (with hair) Hierarchical clustering using this gene signature was performed on a set of nonsyndromic skin) samples. This analysis identified gene clusters that were differentially expressed between these samples, as well as gene clusters that were systemically equivalent across lesioned and non-lesioned samples. We subsequently removed from the first expression set any genes classified into clusters that correlate with lesion versus non-lesion status. This mainly removed a significant number (but not all) of keratin and keratin related proteins from the signature.

  The resulting gene expression signature, alopecia areata gene signature (AAGS), consists of a total of 136 unique genes (Table A), and the entire training cohort is divided into two suitable supers corresponding to the control and disease states. Sufficient information was provided to cluster into clusters (FIG. 24A). Clustering these genes by co-expression also revealed a module of two separate genes, with greater diversity of co-expression in genes up-regulated in disease states (FIG. 24B). As a qualitative measure of genes differentially expressed between affected and unaffected patients, we analyzed them for functional annotation enrichments. This analysis revealed the presence of HLA genes, immune response elements, and inflammation and cell death pathway gene expression in affected patient samples (FIG. 24C). The two most significant superclusters of AAGS were the transmembrane signal peptide (p = 2.8 × 10-11) and the secreted intercellular signal peptide (p = 2.1 × 10-10). As expected, this list also includes several antigen presenting elements and immune response elements associated with AA and autoimmune diseases (FIG. 30). These results indicate that there are significant changes in multiple biological processes in AA presenting cells. We infer that several subsets of these genes arise from the scalp skin and are required to induce invasive mobilization.

  There must also be pre-filtering; otherwise there is significant evidence for immune-related genes derived from infiltrating immune cells that may disrupt the identification of skin-specific molecular programs. Genetic markers associated with immune cells or immune responses were detected as part of AAGS including CD8a, CXCL9 / 10 and CCL5 / 18/20/26. Defining skin-specific molecular behaviors that contribute to AA in primary patient biopsy samples is a difficult task due to the presence of infiltrating T cells and secondary response pathways in AA skin samples.

Utilizing a control network to deconvolute AAGS skin and immune signatures into control modules With a clear definition of disease signatures, we have changed the AAGS skin molecular program from molecular programs derived from infiltrating immune cells, Attempted to deconvolve in a general and unbiased manner. Rather than using GO pathway enrichment or other prediction-based methods that rely on a priori knowledge and potentially ambiguous predictions, we instead map to a skin-specific control network By identifying genes that cannot, we make use of inferred regulatory networks under the hypothesis that non-skin (immune infiltration) gene expression can be filtered.

  A scalp skin transcriptional control network was generated using the ARACNe algorithm and associated software suite (see Experimental Procedures). Specifically, to generate a network, we have normal (unaffected) whole skin biopsy and primary cultured dermal fibroblasts and dermis that contain little or no T cell infiltration. A cohort of 106 primary scalp skin samples consisting of several samples of papillary cells was included. This network represents a control network in non-invasive skin-derived tissue and serves as the basis for deconvolution that occurs in two major steps as detailed in FIG.

  For deconvolution of the control module, the genes in AAGS are mapped directly to the control network (FIG. 25A; see experimental procedure for details). Genes in AAGS are only retained if there is a direct regulatory interaction between it and TF using the control logic of the cutaneous ARACN network (red, solid line edges). Genes uniquely derived from infiltrating immune cells have no significant expression in the ARACNe network and are subsequently removed from the AAGS (black, dotted edge) for the skin and added to the immune gene signature (IGS).

  IGS was used as a “negative control” signature adapted from previous work characterizing cancer immune invasion. The signature was defined as a set of genes that are specifically expressed in each immune cell type including T cells, B cells, mast cells and macrophages. This step iteratively redefines AAGS and IGS by separating those genes whose control can be explained by a non-invasive control network (AAGS) from those that cannot (IGS). By extension, we anticipated that filtered AAGS would be sufficiently enriched in skin gene expression to produce accurate skin-specific regulons.

  As shown in FIG. 25B, 13 invasion-specific genes were removed from AAGS when passed through a skin-specific control network (9.5% of all signatures). These genes are also listed in Table A. This resulted in two mutually exclusive gene modules (no duplicate genes, p = 1.77 × 10 −4), AAGS and IGS. Subsequent pathway enrichment analysis further confirms the loss of statistical enrichment in the “T cell activation” and “immune response” categories, while other clusters containing known skin immune response elements (such as HLA genes). Retained. This left a total of 123 genes in AAGS that we interpret as representing all end organ programs associated with AA pathology, including immune mobilization initiated in end organs and immune responses (Table 1). A, star item).

  Note that we have differentiated between annotations associated with immune cells (eg, CD8a) and annotations associated with immune response genes (eg, HLA). The former is removed by a control network that is not represented in the skin control network. The latter represents a response element in the skin and is associated with the pathology of the disease, so it is a signature gene that the inventor intends to retain.

  Clustering filtered AAGS reveals two distinct molecular modules that define the transition from an unaffected patient (FIG. 25B, second) to an AA disease state (FIG. 25B, FIG. 3) did. Each node represents a gene within the signature, and its size represents the relative expression in each state (larger means higher expression). We have labeled these gene groups: (1) genes whose expression increases when transitioning to a disease state, and (2) genes whose expression is lost in the transition described above. This filtered AAGS reflects the end organ-specific gene module and serves as an input to the MR analysis.

IKZF1 and DLX4 are MR of invasive mobilization by cutaneous AAGS and dilatation The next step is most important in identifying end organ-specific MRs. We performed MR analysis independently and in parallel for both deconvolved AAGS and IGS using the scalp skin control network (FIG. 25C, first, red outline). Using only the regulatory interactions expressed in the skin, we identified the transcriptional regulator with the highest specificity for deconvolved AAGS (red arrow) and IGS (black arrow) The analysis was repeated. This step compares AAGS to IGS for control logic in the scalp skin, as opposed to direct coverage of gene expression. This analysis uses skin-specific molecular control networks to assess which TFs are the best candidates for deconvolved AAGS (not for IGS). We identify skin-specific candidate MRs by retaining only those candidates that are significant in AAGS coverage and not significant in IGS coverage.

  Of the candidate MRs that are specifically significant for AAGS, we have identified a greedy sort (greedy) to identify the minimal number of regulators needed to maximize AAGS coverage. sort) was adopted. We found that two MRs were sufficient to cover> 60% of AAGS: IKZF1 and DLX4. Any additional candidates boosted coverage with a statistically insignificant margin (<5%). We found that the maximum AAGS fidelity (most faithful reproduction of expression signature) and efficiency (minimum regulator required) were these two genes (IKZF1p = 4.17 × 10 −4 and DLX4p = 4). Conclude that it can be achieved through (.8x10-10, FDR modification).

  An equivalent MR analysis performed on the IGS module failed to generate statistically significant or meaningful MRs using the scalp skin control network. Specifically, the best candidates for AAGS, IKZF1 and DLX4 fall statistically irrelevant (falling from 1st and 2nd to 159th and 210th respectively, FDR = 1) (FIG. 25C, IGS .FDR). Performing MR analysis on AAGS without deconvolution cannot be accurately mapped to MR, but nevertheless due to the presence of contaminating genes in the signature that are detrimental to enrichment in the analysis (p-value and MR candidates cannot be generated with typically expected thresholds (in both signature coverage).

  These two candidates use regulatory interactions from a specific tissue context (scalp skin) that are separate from any immune-specific control module that is deconvolved away using this method. Represents the minimum number of control factors required to reproduce AAGS. Thus, IKZF1 and DLX4 represent genetic control modules in the scalp skin that contribute to AA pathogenesis (FIG. 25C, last) and may be sufficient to induce invasion mobilization in an AA-like manner.

  Although it has never been unprecedented that it may have a role in cells outside the immune system, identification of IKZF1 was unexpected because IKZF1 is a well-established T cell differentiation factor. However, this analysis does not imply that MRs like IKZF1 have no role in T cells that contribute to AA pathogenesis; rather, IKZF1 functions additionally in the scalp skin and interacts between tissues. It is important to note that there is significant evidence to mediate the action.

Expression of IKZF1 and DLX4 induces AAGS-like signatures in normal hair follicle papilla and human keratinocytes To validate MR predictions from functional studies, we have determined that IKZF1 and IKZF1 and The sufficiency in overexpressing DLX4 exogenously and affecting the expression of AAGS was tested. We cloned two isoforms, DLX4 and IKZF1, for exogenous expression in cultured cells. The active IKZF1 isoform served as the experimental arm of the study, while the isoform lacking the DNA binding domain was included as a negative control (IKZF1δ). The inventors expressed these genes in cultured primary human hair follicle papilla (huDP) and human keratinocytes (HK). This experimental system allowed us to directly test two separate but related hypotheses: (1) IKZF1 and DLX4 can induce AA-like recruitment of immune cells, (2) They do it through expression in the skin (not immune infiltrates).

  We identified a set of genes that were significantly differentially expressed in the same direction in IKZF1 and DLX4 transfections across both cell types. Unsupervised hierarchical clustering of all samples based on these transcripts reveals clean co-segregation of IKZF1 and DLX4 transfections from IKZF1δ and RFP (red fluorescent protein) controls (FIG. 26A). . In addition, we found that subclustering within these supergroups was not biased based on the cell type used (HK did not cluster with HK and DP did not cluster with DP). And support that we have identified a situation-independent effect of MR overexpression. Interestingly, we observed that DLX4 transfection resulted in increased levels of IKZF1 transcripts and proteins, whereas IKZF1 transfection did not affect DLX4 expression (FIGS. 26B and 26C). .

  Subsequently, the inventors examined expression data for enrichment of AAGS genes using gene set enrichment analysis (GSEA). We performed two differential gene expression studies comparing IKZF1 transfection versus RFP control and DLX4 transfection versus RFP control. The results show that ectopic expression of the MRs is followed by significant enrichment in the induction of AAGS (IKZF1p = 0.122 and DLX4p = 2.08 × 10-4; FIGS. 26D and 26E).

IKZF1 and DLX4 expression is sufficient to induce NKG2D-mediated cytotoxicity in normal cultured skin. It suggests that MRs can mediate. However, the functional relevance of these MRs to autoimmunity and immune invasion is whether their expression is sufficient to induce a targeted autoimmune response. To investigate this ex vivo, we measure the level of cytotoxic cell death in HK and huDP cells when exposed to peripheral blood mononuclear cells (PBMCs). The experiment was performed.

  We re-transfected both HK and huDP cells with one of four expression constructs: IKZF1, DLX4, RFP (negative control) or IKZF1δ (negative control). 24 hours after transfection, these cells were incubated with fresh purified PBMCs. We additionally cultured human dermal fibroblasts and autologous healthy donor PBMCs. PBMCs were obtained from healthy control subjects with no history of AA or other autoimmune diseases.

  In all comparisons, we observed a statistically significant increase in PBMC-dependent cytotoxicity for IKZF1 and DLX4 transfections compared to RFP and IKZF1δ controls (FIG. 27, middle). Pillars, overall bar height). Patient-compatible PBMCs and RFP control transfected fibroblasts showed no evidence of cytotoxic interaction, as expected in healthy target cells (FIG. 27A, middle). However, the introduction of IKZF1 and DLX4 was sufficient to induce interactions between these previously uninteracted cells, resulting in a significant increase in overall cytotoxicity. In a similar manner, both huDP cells (FIG. 27B, center) and HK cells (FIG. 27C, center) showed a significant increase in cytotoxic sensitivity to PBMC above background levels.

  Since we have previously shown that the possible pathogenic immune cells in AA are CD8 + NKG2D + activated T cells, we also have NKG2D blocking antibodies (in order to prevent NKG2D-dependent interactions ( All procedures were performed with the addition of (see Experimental Procedures). In all cases, we observed that NKG2D blockade suppressed cytotoxicity in both IKZF1 and DLX4 treatments to a level comparable to controls (FIG. 27, middle, gray bar). ). From the difference between inhibitor treated and untreated cells, we can infer cytotoxicity that is NKG2D-dependent (FIG. 27, middle, white bar), which is a relative fold change. Can be normalized to what is observed in the control for analysis. From NKG2D blockade, we observed a specifically statistically significant increase in IKZF1 and DLX4 transfections across all studies (FIG. 27 right). Again, it did not show significant cytotoxicity compared to control transfection. Despite a statistically significant but small (<10%) increase in NKG2D-independent cytotoxicity, an approximately 2- to 8-fold increase in NKG2D-dependent cytotoxicity compared to both controls was there. We have shown from these experiments that IKZF1 and DLX4 can induce NKG2D-dependent interactions with normal PBMCs, which demonstrates the toxicity of transfected cells regardless of the exact tissue type. Conclude that will bring.

  Importantly, these experiments are in contrast to infiltrating cells because exogenous modifications were made strictly on normal cultured cells and exposed to healthy PBMCs from a source with no history of autoimmune disease. Specifically, it establishes cell-autonomous function for IKZF1 and DLX4 in scalp skin.

MR expression allows reconstitution of directional skin-specific MR modules for infiltration mobilization After establishing that IKZF1 and DLX4 are sufficient to induce AAGS, we use this data And tried to completely rebuild the AA MR module. Since we can infer that the control is TF Math Eq T, ARACN detects a direct transcriptional dependency between TF and a non-regulated gene that is a potential target (T) be able to. ARACN cannot infer the directional interaction between TF-TF pairs, and due to the control equivalence of TFs, it cannot subsequently infer the secondary T of MRs (FIG. 28A). ,the first). However, since we directly perturbed HKs and huDPs using specific MRs (FIG. 28A, asterisks), we can infer directionality using gene expression data. . If TFB is MR (TFA) T, then overexpression of TFA will result in differential expression of TFB and we can infer that it is TFA Math Eq TFB . Subsequently, any marker gene in the signature associated with TFB can be associated with MR as a secondary T TFA Math Eq TFB Math Eq T (FIG. 28A, top). If TFB functions upstream or parallel to MR, then TFB and T expression would not be affected by TFA overexpression (FIG. 28A, bottom).

  Using this logic, we reconfigured the control module to measure the full range of coverage obtained by overexpressing IKZF1 and DLX4 in these cellular situations. We (1) responded to IKZF1 / DLX4 expression in experiments and (2) any of TFs predicted to have mutual information with the expressed MR by ARACNe to the control module The downstream T was mapped. We found that 78% of responding AAGSs were within 2 ° of downstream separation from MRs IKZF1 and DLX4 based on these criteria (FIG. 28B).

IKZF1 and DLX4 can be used to predict both immune invasion and disease severity in independent cohorts. As a validation of this module, we return to the original AA array cohort and perform machine learning analysis did. We attempted to classify the validation AA set into controls and affected samples using only the estimated IKZF1 and DLX4 activity. Using the initial training set from FIG. 24, we arranged samples in the search space for two dimensions: IKZF1 consensus activity (x-axis) and DLX4 consensus activity (y-axis) (experimental procedure). reference). From the training set, we generated a topographical map of the consensus activity space and defined the extent of IZKF1 and DLX4 activity associated with the control sample, mottled AA and AT / AU samples (FIG. 28C, Black line). The area in FIG. 28C closest to the origin of the plot represents the lowest combined IKZF1 and DLX4 activity; its upper limit (lower black line) maximizes the difference between the control and all AA patients This is a support vector machine (SVM) margin. The next upper limit (upper black line) represents the SVM margin that maximizes the separation of AT / AU patients from AAP.

  Using these measures of MR activity, we looked at the validation set and tested for the predictive power of these parameters in separating patients and controls. In addition to clinical severity, we observed a strong ability to separate samples into disease and control states (FIG. 28C, top, p <1 × 10−5). A centroid map for each patient subgroup provides a clearer picture of how patient group transition from control (NC) to AAP and AT / AU is reflected by relative IKZF1 and DLX4 activity (Figure 28C, bottom). For comparison, we also included centroids for AAP non-lesion sample biopsies not included in the training set.

Deconvolution applied to independent inflammatory skin diseases identifies known genes. For comparison and to provide a proof of concept about the generality of the approach, we have developed atopic dermatitis. Publicly available gene expression data sets for (AD) and psoriasis (Ps) were downloaded. We generated gene expression signatures for each disease by comparing lesion biopsies with unaffected biopsies, similar to AA analysis (FIGS. 29A and 31). Direct comparison of genes within AA, AD and Ps signatures revealed statistically significant evidence that AAGS is distinct from both AD and Ps (p = 0.003 and 3 respectively). .93x10-13). In contrast, comparison of AD and Ps signatures with each other revealed statistically significant evidence for possible shared molecular pathology, with shared molecular signatures and extension (p = 0). .0173).

  These two signatures were applied to the pipeline. FIG. 29B reports the top 5 MRs identified after analysis, ranked by their total coverage of appropriate disease signatures (Ps or AD). Also provided is a rank of MRs using the corresponding deconvoluted IGS. These results indicate that important control hubs associated with AA (especially IKZF1 and DLX4) are unique to AA. Each disease was assigned its own unique list of MRs, but there was also an overlap of two candidate MRs: SMAD2 and HLTF in AD and Ps. SMAD2 and TGFBR1 are TFs with published evidence of involvement in Ps, and the pipeline can identify them without a priori evidence using the basic definition of Ps gene expression signatures. It was. These results demonstrate the effectiveness of modeling complex genetic behavior as a control module to identify pathological mechanisms.

DISCUSSION Gene regulatory networks utilizing genome-wide profiling and systemic generation and analysis of gene expression data have proven to be a tool for the study of complex diseases. An integrated project to investigate functional interactions has recently been utilized for genome-wide expression signature deconvolution and cross-tissue interactions in diabetes and atherosclerosis. Yes. These studies have been invaluable in identifying invasion gene signatures that provide insight into the types of pathogenic immune infiltration associated with disease. These studies also provide significant genomic-level coverage of interacting tissue regulatory activity and tissue-level gene panels, as well as systematic algorithms developed in cancer and Alzheimer's research This helped identify driver genes from eQTL and other genome-related studies.

  However, particularly in contexts such as AA, characterize the modular modulation of individual pathogenic molecular behaviors within gene expression profiles and how they translate into physiological interactions between diseased tissues Little has been done. Modeling physiological traits as a genetic program controlled by MR provides a unique and powerful perspective in the study of complex diseases. This approach directs large gene expression signatures to a relatively small number of selected MRs, which then become T for gene therapy or manipulation via drugs and small molecules.

  Here, we compare complex networks of end-organ (scalp) and infiltrated (immuno-infiltrating) tissues of tissues in AA by comparing regulatory networks of various skin conditions (infiltrating and normal) Extend the application of regulatory networks to examine conditions. In addition to their typical use to identify key control hubs that govern molecular phenotypic switches, we have based on whether they are accurately represented in independent context-dependent networks. Establishing that these networks can be used to isolate and compartmentalize molecular behavior from different tissues. This makes it possible to more accurately identify tissue-specific molecular programs from mixed samples that contribute to integrated interactive physiological behaviors such as immune invasion. Using this pipeline, we were able to reconstruct MR that mediates invasion from the skin as well as the context of AA, but analysis of Ps and AD is generally inflammatory It provides additional candidates for gene regulation of skin diseases and demonstrates the general applicability of this approach.

  Apart from its direct impact on AA pathology, this study provides proof of principle for two important and generalizable concepts: (1) Complex interactions between two tissues are quantifiable molecules Can be modeled as gene expression modules; and (2) these modules and their regulators are extracted from expression data, compartmentalized into tissues, and induce relevant interactions in normal cell types Selected to do. This reproduces AAGS upon ectopic expression of MRs IKZF1 and DLX4, and then uses normal (non-AA) PBMC only through manipulation of IKZF1 and DLX4 expression in the terminal organ itself, using non-AA cell lines Was demonstrated by the ability to induce enhanced cytotoxicity (without genetic manipulation of PBMC).

  Specifically, this analysis identified MRs that were sufficient to induce interactions with immune cells when expressed only in the scalp. Even in patient-adapted situations with samples from healthy AA unaffected patients, IKZF1 and DLX4 expression induces abnormal NKG2D-dependent interactions between dermal fibroblasts and PBMCs, causing cytotoxicity It was enough. These interactions are not present in control transfection, and IKZF1 or DLX4 expression is sufficient to induce interaction with normal immune cells regardless of the particular tissue or host identity. Shown was repeated with two other (non-patient compatible) cell types. Identification of IKZF1 and DLX4 would not have been possible without network-based deconvolution because the significant presence of signatures penetrating the original AAGS prevented any accurate identification of candidate MRs. Instead, network-based deconvolution has identified MRs that can induce specific molecular interactions in any of several molecular situations that are completely independent of AA itself.

  Since IKZF1 has been extensively studied in the context of T cell differentiation, the identification of IKZF1 was unexpected. However, the identification was only obtained from using deconvolved AA signatures, not IGS, using skin-derived regulatory logic. If we have relied on public databases, previous literature, or GO annotations to filter gene expression data, we have for extensive annotation as T cell differentiation factors. IKZF1 would have been completely ignored and removed. Instead, by looking to the regulatory network, we have found that local expression of IKZF1 may have pathogenicity relevance, regardless of its established role in immune cells. The possibility of being able to be identified.

  Although IKZF1 is well characterized in the context of immune cells, the role of IKZF1 outside immune cells is not unprecedented in the literature. Loss of the IKZF1 and DLX4 loci is also associated with tumorigenesis and low-grade squamous intraepithelial lesions in colorectal cancer, lung cancer and breast cancer. These studies have obtained genomic information directly from the tumor mass and show that somatic losses at these two loci can contribute to cancer pathophysiology as a genomic modification of the final organ. Studies of IKZF1 and DLX4 as MRs that induce immune invasion support these results and increase the likelihood that loss of these loci contributes to immune evasion in cancer. Furthermore, these observations, and the identification of IKZF1 and DLX4 as MRs for immune infiltration recruitment, support the function of IKZF1 in addition to its role as T cell-specific differentiation factors, and autoimmunity in AA and We present a hypothetical indication that tumor immune evasion exists at the extreme opposite of normal immune interactions. Loss of MR in immune infiltration is associated with cancer, and its overexpression is associated with the development of autoimmune disease in AA.

  We have system biology and network analysis to model the molecular mechanisms that mediate the interaction between two different tissues, to identify key regulators, and in other contexts Shown that it can be used to use them to re-create. Although the output for validation of these MRs was ultimately the induction of cell death, the function of these MRs in the context of autoimmune disease ultimately signals immune infiltrates and It is to induce a molecular profile that replenishes things. So far, the application of system biology has been mainly to identify “breakpoints” in the cell-autonomous molecular behavior of cancer. Controlled induction of cross-tissue interactions, particularly those involving the immune system, can potentially lead to a potentially significant path for modeling complex genetic traits with regulatory networks not previously possible. Invite. We provide a proof-of-concept framework that can be used to actively partition molecular behavior for research, even in complex diseases involving interactions between different tissues.

Experimental Procedures This section describes the less common or unique methods implemented in this study.

ARACNe
To generate a context-specific transcriptional interaction network for the scalp, we used the ARACNe algorithm in a set of 128 microarray experiments independent of the analytical cohort in this study. These experiments show a platform fit (Affymetrix U133 2plus) obtained for whole skin samples from normal whole skin biopsies, AA patient biopsies, microdissected dermal papillae, and a mixture of separated dermis and epidermis samples. Represents data. These samples collectively provide the heterogeneity necessary for accurate detection of transcriptional dependence in the scalp. Experiments were pooled, worked up as described above, and standard ARACNe analyzes were performed. The ARACNe software suite is available from the Califano laboratory website (http://wiki.c2b2.columbia.edu/califanalab/index.php/Software).

MR analysis MRs for specific gene expression signatures were defined as TFs in which the direct ARACN prediction T (regulon) was statistically enriched in gene expression signatures. Each TF regulon was tested for AAGS enrichment using the Fisher's exact test, FDR = 0.05. This analysis allows ranking and determination of the minimum number of TFs required to specifically cover gene expression modules associated with physiological traits. http: // wiki. c2b2. columnia. edu / califanolab / index. php / Software / MARINA

MR Activity Classification Indicators The ARACN prediction T of IKZF1 and DLX4 was integrated with exogenous gene expression studies to identify all genes of AAGS that could be mapped as T of IKZF1 and DLX4. This was done by crossing AAGS with the ARACN regulon of IKZF1 and DLX4. These two sets of intersections were then screened in expression studies of any gene that responded with at least 25% fold change. This gene set was used to construct a consensus “metaactivity” for the IKZF1 and DLX4 loci. The rank-normalized change of each gene across the AA patient cohort was integrated into the mean as a consensus measure of the relative activity of the parent MR.

  These values were then used to define a 2D search space, Math Eq (X = IKZF1 metaactivity and Y = DLX4 metaactivity) to classify each of the patients in the AA training set. The meta-activity vectors (vectors) were rank transformed so that the minimum value was coupled to the origin (0,0) of the search space and the activity index was positive. This transformation has no effect on the results other than projecting the search space onto a more intuitive grid for display purposes. Both axes are coupled between [0, n], where n is a positive number.

The classification in this space was done using a modified non-linear soft margin SVM algorithm. The algorithm is formulated. :

  This algorithm defines a vector set Math Eq that exists in the search space Math Eq, so that each given pair Math Eq maximizes the likelihood ratio Math Eq. This function is defined such that Math Eq is the next order of disease severity relative to Math Eq, and Math Eq and Math Eq are quadrants I and III of the grid created by the hyperplanes Math Eq and Math Eq. It is. Samples in the training set are mapped to each grid with known molecular subtypes, and likelihood ratios are calculated for subtype separation as defined by Math Eq. The severity ranking used for Math Eq was Normal <Mild <Severe. Thus, each coordinate set in Math Eq defines a point to a nonlinear plane that maximizes the separation between samples of different molecular classes in the IKZF1 / DLX4 metaactivity space.

Cytotoxicity assay PBMC-dependent cytotoxicity was measured using the CytoTox 96 non-radioactive cytotoxicity assay available via Promega. For sample and solution processing, we followed the manufacturer's protocol. Optimization of PBMC: T was performed as follows, but using variable concentrations (1: 1, 5: 1 and 10: 1) (FIG. 32).

  Cytotoxicity experiments were performed in a 96 well format and each treatment was performed in triplicate. Transfection was performed 36 hours before the experiment. On the day of the experiment, HK and huDP cells were trypsinized and diluted with Dulbecco's modified Eagle's medium (DMEM) to a working stock. The T concentration per well was 80,000 cells in 50 μL DMEM and combined with 800,000 PBMCs. The NKG2D inhibitor was a human NKG2D MAb (clone 149810) from R & D Systems (Cat. MAB139) used at a final concentration of 20 μg / mL. Each transfection was assigned in triplicate according to the manufacturer's instructions.

Gene Expression Studies A total of 122 samples from 96 patients were profiled with an Affymetrix U133 2Plus array consisting of 28 AAP patients, 32 AT / AU patients, and 36 unaffected controls. The remaining 26 samples correspond to patient matched non-lesion biopsies from the AAP cohort. These non-lesion samples were not included in the initial signature inference, but were used later (below). RNA from these patient biopsies was isolated and processed with an Affymetrix U133 2Plus array. Post-processing of the data was done via R using MAS5 normalization with a standard package available via Bioconductor. These data are available as GSE68801 at Gene Expression Omnibus. This data set was divided into two sets for training and validation.

  The first panel of genetic markers was identified by two different expression analyzes comparing (1) AA versus unaffected and (2) lesion versus non-lesion in the training set. The threshold was set for differential expression with p <0.05 and fold change> 25%. This relatively slow threshold was implemented because network analysis is based on consensus. The analysis is not primarily related to the candidate rank, but rather relies on having sufficient molecular information to infer TF activity. This approach applies noise to the entire data set and is normalized from consensus in both ARACNe and master regulator analysis (see below), so noise that may be introduced by a more relaxed threshold Inevitably strong against. All X-linked and Y-linked genes were further removed to remove any possible gender bias in ranking and clustering of differentially expressed genes.

Gene Set Enrichment Analysis (GSEA)
GSEA is a method for measuring non-parametric statistical enrichment in expression differences in a defined gene panel. A default differential expression analysis between the experimental and control cohorts is performed, and the genes are ranked by thresholdless expression differential (all genes are included). This can be done according to any criteria (fold change, p-value, etc.) specified by the user.

  This enrichment score is then compared to an empirically generated null distribution by shuffling the sample label, i.e., randomizing the case and control samples and repeating the analysis. . This can be repeated over 1000 iterations to generate a null distribution of enrichment scores and the observed scores can be compared to generate a p-value.

Cloning Each primer pair provided below was used in a PCR reaction with AccuPrime Taq PCR mix according to the manufacturer's protocol for cDNA derived from HEK293T cells. CDNA was generated from cultured cells using the SuperScript First-Strand Synthesis System from Invitrogen. The PCR product was used up in gel electrophoresis and any isoform present was excised separately using Qiagen Gel Extraction Kit.

  MRNA fidelity was confirmed via sequencing from Genewiz and digested with the appropriate enzymes (SPEI and ASCI) from New England Biosystems in SmartCut buffer for 2 hours. The pLOC-RFP vector was digested in parallel and the cleaved backbone was excised by gel extraction. After purification of the backbone and insert, each insert was ligated into the cut pLOC vector using a RapidLigation Kit from Roche according to the manufacturer's protocol and transformed into DH5α cells for amplification.

  Successful transformations were verified for sequence fidelity via colony PCR and sequencing (Genewiz). The correct construct was amplified and purified for experimentation by Maxiprep (Qiagen).

  The primers used to clone the gene for insertion into the pLOC vector are provided below in the following format, 5 'to 3': spacer-enzyme-mRNA sequence.

IKZF1.1
Forward GGC-ACTAGT-ATGGATGCTGATGAGGGTCAA
Reverse ATT-GGCGCGCC-TTAGCTCATGTGGAAGCGGT
IKZF1.2
Forward GGC-ACTAGT-ATGGATGCTGATGAGGGTCAAG
Reverse ATT-GGCGCGCC-TTAGCTCATGTGGAAGCGGT (same as 1.1)
DLX4
Forward GGC-ACTAGT-ATGAAACTGTCCGTCCTACCCCC
Reverse ATT-GGCGCGCC-TCATTCACACGCTGGGGGCTGG

Cell culture and transfection Both huDP and HK cells were maintained in DMEM 10% FBS at 37 ° C. and 5% CO 2 in standard conditions for growth. HuDP cells are cultured primary human dermal papilla microdissected from human skin samples. In this study experiment, only huDP and HK cells with passage number <6 (<6) were used.

  Cells were transformed with the pLOC expression construct using JetPRIME transfection reagent according to the manufacturer's protocol. Transfections were performed overnight using a 2: 1 concentration of reagent (ul) and DNA (ug).

MR Rescue Microarray Transfection of IKZF1 and DLX4 into HK and huDP cells was performed in cells cultured in 10 cm plates as described above. After 36 hours post-transfection, the cells are harvested in PBS with scraped cells, and then purified RNA is lysed and processed using Qiagen's RNeasy kit according to the manufacturer's protocol. did. RNA quality control was performed using a spectrometer and was submitted for processing on an Affymetrix human U133 2Plus array by the Columbia facility (Department of Pathology). Array data was normalized again and processed using MAS5 normalization via the Bioconductor package in R.

qPCRs
8 primary lesional biopsies (one found to be degraded and excluded from the study), 4 unaffected controls, and 5 pairs of patient-matched and non-lesioned samples Quantitative PCR reactions were performed on cDNA extracted from these independent cohorts. A reaction mix using SYBR Green was made in a 25 uL volume according to the manufacturer's protocol and analyzed on a 7300 series real-time PCR instrument from Applied Biosystems. Primers for each gene are provided at the end of this section.

  All samples were tested in a technical triplicate in a stamp plate format (each replicate was prepared in one plate with all samples and controls prepared at once and repeated three times). Data from these replicates were analyzed by the δδCT method and all experimental series were normalized to the mean normalized value of the control tissue. The standard error of the mean (SEM) was obtained across comparisons using standard statistical error propagation.

  Primers for assaying transcripts by qPCR are provided below 5 'to 3'. Primers for full length amplification of DLX4 were used because the transcript is approximately 300 bp (the optimal transcript length for the protocol provided is 200-300 bp).

IKZF1
Forward ACTCCGTTGGTAAACCTCAC
Reverse CTGATCCTATTCTTGCACAGGTC
DLX4
* Same as cloning primer *
ACTB
Forward GAAGGATTCCTATGTGGGCGAC
Reverse GGGTCATCTTCTCCGCGTGTG

Separation of Fresh Peripheral Blood Mononuclear Cells (PBMC) Fresh PBMC were isolated from whole blood drawn in the evening prior to the intended cytotoxicity assay. Whole blood using Histopaque-1077 reagent (Ficoll) by diluting an 8 mL aliquot of 1: 1 whole blood in sterile PBS and overlaying the solution onto Ficoll in a final volume ratio of 2: 1. Separated from. This solution was centrifuged at 1200 rpm for 45 minutes. The monocyte-containing interfacial layer was isolated, diluted with 5 volumes of sterile PBS, and centrifuged again at 1500 rpm for 15 minutes. The supernatant was discarded and the pellet was resuspended in 3 ml DMEM 10% FBS. Cell counts were performed using a hemocytometer and the solution was diluted with DMEM 10% FBS to a final concentration of 1 × 10 6 cells per mL. This was stored overnight at 37 ° C. and 5% CO 2 for the next morning experiment.

  In addition to the various embodiments shown and claimed, the disclosed subject matter also relates to other embodiments having other combinations of features disclosed and claimed herein. As such, certain features presented herein may be other features within the scope of the disclosed subject matter, such that the disclosed subject matter includes any suitable combination of features disclosed herein. Can be combined with each other in a way. The foregoing description of specific embodiments of the disclosed subject matter has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the disclosed subject matter to the disclosed embodiments.

Various publications, patents and patent applications, and protocols are cited herein, the contents of which are hereby incorporated by reference in their entirety.

Claims (26)

A method of treating alopecia areata (AA) in a subject comprising:
(A) identifying the subject's AA disease severity by detecting a biomarker indicative of the disease severity; and (b) applying a therapeutic intervention appropriate to the identified disease severity. The steps to
Including a method. A method of treating alopecia areata (AA) in a subject comprising:
(A) identifying a tendency of a subject with AA to respond to JAK inhibitor treatment by detecting a biomarker indicative of the trend; and (b) identifying a tendency of the subject to respond to the inhibitor. Administering a JAK inhibitor to the subject when the indicated biomarker indicates,
Including a method. A method of treating alopecia areata (AA) in a subject comprising:
(A) administering a JAK inhibitor to the subject; and (b) detecting a biomarker indicative of responsiveness to JAK inhibitor treatment; and
(C) Thereafter, the response by either (1) continuing the administration of the JAK inhibitor, (2) changing the administration of the JAK inhibitor, or (3) discontinuing the administration of the JAK inhibitor. Adapting the administration of said JAK inhibitor based on
Including a method.   Detecting the biomarker is performed on a sample obtained from the subject, the sample comprising skin, blood, serum, plasma, urine, saliva, sputum, mucus, semen, amniotic fluid, mouthwash and 4. The method according to any one of claims 1 to 3, wherein the method is selected from the group consisting of bronchial lavage fluid.   The method of claim 4, wherein the subject is a human.   The method of claim 4, wherein the sample is a skin sample.   The method of claim 4, wherein the sample is a serum sample.   The method of claim 4, wherein the biomarker is a gene expression signature.   9. The method of claim 8, wherein the gene expression signature comprises one or more gene expression information from the following group of genes: KRT-related genes, CTL-related genes and IFN-related genes.   The method according to claim 9, wherein the KRT-related genes include DSG4, HOXC31, KRT31, KRT32, KRT33B, KRT82, PKP1 and PKP2.   The method according to claim 9, wherein the CTL-related genes include CD8A, GZMB, ICOS and PRF1.   The method according to claim 9, wherein the IFN-related genes include CXCL9, CXCL10, CXCL11, STAT1 and MX1.   9. The method of claim 8, wherein the gene expression signature is an alopecia areata disease activity factor (ALADIN).   9. The method of claim 8, wherein the gene expression signature is an alopecia areata gene signature (AAGS) comprising one or more genes as defined in Table A.   15. The method of claim 14, wherein the gene expression signature is IKZF1, DLX4, or a combination thereof.   9. The method of claim 8, wherein the detection is performed by a nucleic acid hybridization assay.   The method of claim 8, wherein the detection is performed by microarray analysis.   9. The method of claim 8, wherein the detection is performed by polymerase chain reaction (PCR) or nucleic acid sequencing.   The method of claim 4, wherein the biomarker is a protein.   20. The method of claim 19, wherein the presence of the protein is detected using a reagent that specifically binds to the protein.   21. The method of claim 20, wherein the reagent is a monoclonal antibody or antigen binding fragment thereof, or a polyclonal antibody or antigen binding fragment thereof.   21. The method of claim 20, wherein the presence of the protein is detected by enzyme linked immunosorbent assay (ELISA), immunofluorescence measurement, or Western blot assay. A kit for treating alopecia areata (AA) in a subject comprising:
(A) one or more detection reagents useful for detecting a biomarker indicative of the disease severity of the subject, and (b) one or more treatments useful for treating AA. reagent,
Including a kit. A kit for treating alopecia areata (AA) in a subject comprising:
(A) one or more detection reagents useful for detecting a biomarker of the subject that exhibits a tendency to respond to treatment reagents useful for treating AA; and (b) treating AA. One or more treatment reagents useful for
Including a kit.   25. A kit according to claim 23 or claim 24, further comprising one or more probe sets, arrays / microarrays, biomarker specific antibodies and / or beads. 25. A kit according to claim 23 or claim 24, further comprising a set of instructions for use of the kit.

Images